feat: use jitsi file system

- Resolved conflicts in server/reflector/views/rooms.py to keep platform-agnostic approach
- Resolved conflicts in www/app/[roomName]/page.tsx to keep VideoPlatformEmbed approach
- Accepted main's version of generated API files (schemas.gen.ts, services.gen.ts, types.gen.ts)
- Removed config-template.ts as per main branch changes

.github/workflows/db_migrations.yml

@@ -2,6 +2,8 @@ name: Test Database Migrations
 
 on:
   push:
+    branches:
+      - main
     paths:
       - "server/migrations/**"
       - "server/reflector/db/**"
@@ -17,6 +19,9 @@ on:
 jobs:
   test-migrations:
     runs-on: ubuntu-latest
+    concurrency:
+      group: db-ubuntu-latest-${{ github.ref }}
+      cancel-in-progress: true
     services:
       postgres:
         image: postgres:17

.github/workflows/test_next_server.yml

@@ -0,0 +1,45 @@
+name: Test Next Server
+
+on:
+  pull_request:
+    paths:
+      - "www/**"
+  push:
+    branches:
+      - main
+    paths:
+      - "www/**"
+
+jobs:
+  test-next-server:
+    runs-on: ubuntu-latest
+
+    defaults:
+      run:
+        working-directory: ./www
+
+    steps:
+    - uses: actions/checkout@v4
+
+    - name: Setup Node.js
+      uses: actions/setup-node@v4
+      with:
+        node-version: '20'
+
+    - name: Install pnpm
+      uses: pnpm/action-setup@v4
+      with:
+        version: 8
+
+    - name: Setup Node.js cache
+      uses: actions/setup-node@v4
+      with:
+        node-version: '20'
+        cache: 'pnpm'
+        cache-dependency-path: './www/pnpm-lock.yaml'
+
+    - name: Install dependencies
+      run: pnpm install
+
+    - name: Run tests
+      run: pnpm test

.github/workflows/test_server.yml

@@ -5,12 +5,17 @@ on:
     paths:
       - "server/**"
   push:
+    branches:
+      - main
     paths:
       - "server/**"
 
 jobs:
   pytest:
     runs-on: ubuntu-latest
+    concurrency:
+      group: pytest-${{ github.ref }}
+      cancel-in-progress: true
     services:
       redis:
         image: redis:6
@@ -30,6 +35,9 @@ jobs:
 
   docker-amd64:
     runs-on: linux-amd64
+    concurrency:
+      group: docker-amd64-${{ github.ref }}
+      cancel-in-progress: true
     steps:
       - uses: actions/checkout@v4
       - name: Set up Docker Buildx
@@ -45,6 +53,9 @@ jobs:
 
   docker-arm64:
     runs-on: linux-arm64
+    concurrency:
+      group: docker-arm64-${{ github.ref }}
+      cancel-in-progress: true
     steps:
       - uses: actions/checkout@v4
       - name: Set up Docker Buildx

.gitignore

@@ -14,4 +14,7 @@ data/
 www/REFACTOR.md
 www/reload-frontend
 server/test.sqlite
-CLAUDE.local.md
+CLAUDE.local.md
+www/.env.development
+www/.env.production
+.playwright-mcp

.gitleaksignore

@@ -0,0 +1 @@
+b9d891d3424f371642cb032ecfd0e2564470a72c:server/tests/test_transcripts_recording_deletion.py:generic-api-key:15

.pre-commit-config.yaml

@@ -27,3 +27,8 @@ repos:
         files: ^server/
       - id: ruff-format
         files: ^server/
+
+  - repo: https://github.com/gitleaks/gitleaks
+    rev: v8.28.0
+    hooks:
+      - id: gitleaks

CHANGELOG.md

@@ -1,5 +1,74 @@
 # Changelog
 
+## [0.10.0](https://github.com/Monadical-SAS/reflector/compare/v0.9.0...v0.10.0) (2025-09-11)
+
+
+### Features
+
+* replace nextjs-config with environment variables ([#632](https://github.com/Monadical-SAS/reflector/issues/632)) ([369ecdf](https://github.com/Monadical-SAS/reflector/commit/369ecdff13f3862d926a9c0b87df52c9d94c4dde))
+
+
+### Bug Fixes
+
+* anonymous users transcript permissions ([#621](https://github.com/Monadical-SAS/reflector/issues/621)) ([f81fe99](https://github.com/Monadical-SAS/reflector/commit/f81fe9948a9237b3e0001b2d8ca84f54d76878f9))
+* auth post ([#624](https://github.com/Monadical-SAS/reflector/issues/624)) ([cde99ca](https://github.com/Monadical-SAS/reflector/commit/cde99ca2716f84ba26798f289047732f0448742e))
+* auth post ([#626](https://github.com/Monadical-SAS/reflector/issues/626)) ([3b85ff3](https://github.com/Monadical-SAS/reflector/commit/3b85ff3bdf4fb053b103070646811bc990c0e70a))
+* auth post ([#627](https://github.com/Monadical-SAS/reflector/issues/627)) ([962038e](https://github.com/Monadical-SAS/reflector/commit/962038ee3f2a555dc3c03856be0e4409456e0996))
+* missing follow_redirects=True on modal endpoint ([#630](https://github.com/Monadical-SAS/reflector/issues/630)) ([fc363bd](https://github.com/Monadical-SAS/reflector/commit/fc363bd49b17b075e64f9186e5e0185abc325ea7))
+* sync backend and frontend token refresh logic ([#614](https://github.com/Monadical-SAS/reflector/issues/614)) ([5a5b323](https://github.com/Monadical-SAS/reflector/commit/5a5b3233820df9536da75e87ce6184a983d4713a))
+
+## [0.9.0](https://github.com/Monadical-SAS/reflector/compare/v0.8.2...v0.9.0) (2025-09-06)
+
+
+### Features
+
+* frontend openapi react query ([#606](https://github.com/Monadical-SAS/reflector/issues/606)) ([c4d2825](https://github.com/Monadical-SAS/reflector/commit/c4d2825c81f81ad8835629fbf6ea8c7383f8c31b))
+
+
+### Bug Fixes
+
+* align whisper transcriber api with parakeet ([#602](https://github.com/Monadical-SAS/reflector/issues/602)) ([0663700](https://github.com/Monadical-SAS/reflector/commit/0663700a615a4af69a03c96c410f049e23ec9443))
+* kv use tls explicit ([#610](https://github.com/Monadical-SAS/reflector/issues/610)) ([08d88ec](https://github.com/Monadical-SAS/reflector/commit/08d88ec349f38b0d13e0fa4cb73486c8dfd31836))
+* source kind for file processing ([#601](https://github.com/Monadical-SAS/reflector/issues/601)) ([dc82f8b](https://github.com/Monadical-SAS/reflector/commit/dc82f8bb3bdf3ab3d4088e592a30fd63907319e1))
+* token refresh locking ([#613](https://github.com/Monadical-SAS/reflector/issues/613)) ([7f5a4c9](https://github.com/Monadical-SAS/reflector/commit/7f5a4c9ddc7fd098860c8bdda2ca3b57f63ded2f))
+
+## [0.8.2](https://github.com/Monadical-SAS/reflector/compare/v0.8.1...v0.8.2) (2025-08-29)
+
+
+### Bug Fixes
+
+* search-logspam ([#593](https://github.com/Monadical-SAS/reflector/issues/593)) ([695d1a9](https://github.com/Monadical-SAS/reflector/commit/695d1a957d4cd862753049f9beed88836cabd5ab))
+
+## [0.8.1](https://github.com/Monadical-SAS/reflector/compare/v0.8.0...v0.8.1) (2025-08-29)
+
+
+### Bug Fixes
+
+* make webhook secret/url allowing null ([#590](https://github.com/Monadical-SAS/reflector/issues/590)) ([84a3812](https://github.com/Monadical-SAS/reflector/commit/84a381220bc606231d08d6f71d4babc818fa3c75))
+
+## [0.8.0](https://github.com/Monadical-SAS/reflector/compare/v0.7.3...v0.8.0) (2025-08-29)
+
+
+### Features
+
+* **cleanup:** add automatic data retention for public instances ([#574](https://github.com/Monadical-SAS/reflector/issues/574)) ([6f0c7c1](https://github.com/Monadical-SAS/reflector/commit/6f0c7c1a5e751713366886c8e764c2009e12ba72))
+* **rooms:** add webhook for transcript completion ([#578](https://github.com/Monadical-SAS/reflector/issues/578)) ([88ed7cf](https://github.com/Monadical-SAS/reflector/commit/88ed7cfa7804794b9b54cad4c3facc8a98cf85fd))
+
+
+### Bug Fixes
+
+* file pipeline status reporting and websocket updates ([#589](https://github.com/Monadical-SAS/reflector/issues/589)) ([9dfd769](https://github.com/Monadical-SAS/reflector/commit/9dfd76996f851cc52be54feea078adbc0816dc57))
+* Igor/evaluation ([#575](https://github.com/Monadical-SAS/reflector/issues/575)) ([124ce03](https://github.com/Monadical-SAS/reflector/commit/124ce03bf86044c18313d27228a25da4bc20c9c5))
+* optimize parakeet transcription batching algorithm ([#577](https://github.com/Monadical-SAS/reflector/issues/577)) ([7030e0f](https://github.com/Monadical-SAS/reflector/commit/7030e0f23649a8cf6c1eb6d5889684a41ce849ec))
+
+## [0.7.3](https://github.com/Monadical-SAS/reflector/compare/v0.7.2...v0.7.3) (2025-08-22)
+
+
+### Bug Fixes
+
+* cleaned repo, and get git-leaks clean ([359280d](https://github.com/Monadical-SAS/reflector/commit/359280dd340433ba4402ed69034094884c825e67))
+* restore previous behavior on live pipeline + audio downscaler ([#561](https://github.com/Monadical-SAS/reflector/issues/561)) ([9265d20](https://github.com/Monadical-SAS/reflector/commit/9265d201b590d23c628c5f19251b70f473859043))
+
 ## [0.7.2](https://github.com/Monadical-SAS/reflector/compare/v0.7.1...v0.7.2) (2025-08-21)
 
 

CLAUDE.md

@@ -66,7 +66,6 @@ pnpm install
 
 # Copy configuration templates
 cp .env_template .env
-cp config-template.ts config.ts
 ```
 
 **Development:**

README.md

@@ -1,43 +1,60 @@
 <div align="center">
+<img width="100" alt="image" src="https://github.com/user-attachments/assets/66fb367b-2c89-4516-9912-f47ac59c6a7f"/>
 
 # Reflector
 
-Reflector Audio Management and Analysis is a cutting-edge web application under development by Monadical. It utilizes AI to record meetings, providing a permanent record with transcripts, translations, and automated summaries.
+Reflector is an AI-powered audio transcription and meeting analysis platform that provides real-time transcription, speaker diarization, translation and summarization for audio content and live meetings. It works 100% with local models (whisper/parakeet, pyannote, seamless-m4t, and your local llm like phi-4).
 
-[![Tests](https://github.com/monadical-sas/reflector/actions/workflows/pytests.yml/badge.svg?branch=main&event=push)](https://github.com/monadical-sas/reflector/actions/workflows/pytests.yml)
+[![Tests](https://github.com/monadical-sas/reflector/actions/workflows/test_server.yml/badge.svg?branch=main&event=push)](https://github.com/monadical-sas/reflector/actions/workflows/test_server.yml)
 [![License: MIT](https://img.shields.io/badge/license-MIT-green.svg)](https://opensource.org/licenses/MIT)
 </div>
-
-## Screenshots
+</div>
 <table>
   <tr>
     <td>
-      <a href="https://github.com/user-attachments/assets/3a976930-56c1-47ef-8c76-55d3864309e3">
-        <img width="700" alt="image" src="https://github.com/user-attachments/assets/3a976930-56c1-47ef-8c76-55d3864309e3" />
+      <a href="https://github.com/user-attachments/assets/21f5597c-2930-4899-a154-f7bd61a59e97">
+        <img width="700" alt="image" src="https://github.com/user-attachments/assets/21f5597c-2930-4899-a154-f7bd61a59e97" />
       </a>
     </td>
     <td>
-      <a href="https://github.com/user-attachments/assets/bfe3bde3-08af-4426-a9a1-11ad5cd63b33">
-        <img width="700" alt="image" src="https://github.com/user-attachments/assets/bfe3bde3-08af-4426-a9a1-11ad5cd63b33" />
+      <a href="https://github.com/user-attachments/assets/f6b9399a-5e51-4bae-b807-59128d0a940c">
+        <img width="700" alt="image" src="https://github.com/user-attachments/assets/f6b9399a-5e51-4bae-b807-59128d0a940c" />
       </a>
     </td>
     <td>
-      <a href="https://github.com/user-attachments/assets/7b60c9d0-efe4-474f-a27b-ea13bd0fabdc">
-        <img width="700" alt="image" src="https://github.com/user-attachments/assets/7b60c9d0-efe4-474f-a27b-ea13bd0fabdc" />
+      <a href="https://github.com/user-attachments/assets/a42ce460-c1fd-4489-a995-270516193897">
+        <img width="700" alt="image" src="https://github.com/user-attachments/assets/a42ce460-c1fd-4489-a995-270516193897" />
+      </a>
+    </td>
+    <td>
+      <a href="https://github.com/user-attachments/assets/21929f6d-c309-42fe-9c11-f1299e50fbd4">
+        <img width="700" alt="image" src="https://github.com/user-attachments/assets/21929f6d-c309-42fe-9c11-f1299e50fbd4" />
       </a>
     </td>
   </tr>
 </table>
 
+## What is Reflector?
+
+Reflector is a web application that utilizes local models to process audio content, providing:
+
+- **Real-time Transcription**: Convert speech to text using [Whisper](https://github.com/openai/whisper) (multi-language) or [Parakeet](https://huggingface.co/nvidia/parakeet-tdt-0.6b-v2) (English) models
+- **Speaker Diarization**: Identify and label different speakers using [Pyannote](https://github.com/pyannote/pyannote-audio) 3.1
+- **Live Translation**: Translate audio content in real-time to many languages with [Facebook Seamless-M4T](https://github.com/facebookresearch/seamless_communication)
+- **Topic Detection & Summarization**: Extract key topics and generate concise summaries using LLMs
+- **Meeting Recording**: Create permanent records of meetings with searchable transcripts
+
+Currently we provide [modal.com](https://modal.com/) gpu template to deploy.
+
 ## Background
 
 The project architecture consists of three primary components:
 
-- **Front-End**: NextJS React project hosted on Vercel, located in `www/`.
 - **Back-End**: Python server that offers an API and data persistence, found in `server/`.
-- **GPU implementation**: Providing services such as speech-to-text transcription, topic generation, automated summaries, and translations. Most reliable option is Modal deployment
+- **Front-End**: NextJS React project hosted on Vercel, located in `www/`.
+- **GPU implementation**: Providing services such as speech-to-text transcription, topic generation, automated summaries, and translations.
 
-It also uses authentik for authentication if activated, and Vercel for deployment and configuration of the front-end.
+It also uses authentik for authentication if activated.
 
 ## Contribution Guidelines
 
@@ -72,6 +89,8 @@ Note: We currently do not have instructions for Windows users.
 
 ## Installation
 
+*Note: we're working toward better installation, theses instructions are not accurate for now*
+
 ### Frontend
 
 Start with `cd www`.
@@ -80,11 +99,10 @@ Start with `cd www`.
 
 ```bash
 pnpm install
-cp .env_template .env
-cp config-template.ts config.ts
+cp .env.example .env
 ```
 
-Then, fill in the environment variables in `.env` and the configuration in `config.ts` as needed. If you are unsure on how to proceed, ask in Zulip.
+Then, fill in the environment variables in `.env` as needed. If you are unsure on how to proceed, ask in Zulip.
 
 **Run in development mode**
 
@@ -149,3 +167,34 @@ You can manually process an audio file by calling the process tool:
 ```bash
 uv run python -m reflector.tools.process path/to/audio.wav
 ```
+
+
+## Feature Flags
+
+Reflector uses environment variable-based feature flags to control application functionality. These flags allow you to enable or disable features without code changes.
+
+### Available Feature Flags
+
+| Feature Flag | Environment Variable |
+|-------------|---------------------|
+| `requireLogin` | `NEXT_PUBLIC_FEATURE_REQUIRE_LOGIN` |
+| `privacy` | `NEXT_PUBLIC_FEATURE_PRIVACY` |
+| `browse` | `NEXT_PUBLIC_FEATURE_BROWSE` |
+| `sendToZulip` | `NEXT_PUBLIC_FEATURE_SEND_TO_ZULIP` |
+| `rooms` | `NEXT_PUBLIC_FEATURE_ROOMS` |
+
+### Setting Feature Flags
+
+Feature flags are controlled via environment variables using the pattern `NEXT_PUBLIC_FEATURE_{FEATURE_NAME}` where `{FEATURE_NAME}` is the SCREAMING_SNAKE_CASE version of the feature name.
+
+**Examples:**
+```bash
+# Enable user authentication requirement
+NEXT_PUBLIC_FEATURE_REQUIRE_LOGIN=true
+
+# Disable browse functionality
+NEXT_PUBLIC_FEATURE_BROWSE=false
+
+# Enable Zulip integration
+NEXT_PUBLIC_FEATURE_SEND_TO_ZULIP=true
+```

compose.yml

@@ -6,6 +6,7 @@ services:
       - 1250:1250
     volumes:
       - ./server/:/app/
+      - /app/.venv
     env_file:
       - ./server/.env
     environment:
@@ -16,6 +17,7 @@ services:
       context: server
     volumes:
       - ./server/:/app/
+      - /app/.venv
     env_file:
       - ./server/.env
     environment:
@@ -26,6 +28,7 @@ services:
       context: server
     volumes:
       - ./server/:/app/
+      - /app/.venv
     env_file:
       - ./server/.env
     environment:

docs/jitsi.md

@@ -0,0 +1,369 @@
+# Jitsi Integration for Reflector
+
+This document contains research and planning notes for integrating Jitsi Meet as a replacement for Whereby in Reflector.
+
+## Overview
+
+Jitsi Meet is an open-source video conferencing solution that can replace Whereby in Reflector, providing:
+- Cost reduction (no per-minute charges)
+- Direct recording access via Jibri
+- Real-time event webhooks
+- Full customization and control
+
+## Current Whereby Integration Analysis
+
+### Architecture
+1. **Room Creation**: User creates a "room" template in Reflector DB with settings
+2. **Meeting Creation**: `/rooms/{room_name}/meeting` endpoint calls Whereby API to create meeting
+3. **Recording**: Whereby handles recording automatically to S3 bucket
+4. **Webhooks**: Whereby sends events for participant tracking
+
+### Database Structure
+```python
+# Room = Template/Configuration
+class Room:
+    id, name, user_id
+    recording_type, recording_trigger  # cloud, automatic-2nd-participant
+    webhook_url, webhook_secret
+
+# Meeting = Actual Whereby Meeting Instance
+class Meeting:
+    id  # Whereby meetingId
+    room_name  # Generated by Whereby
+    room_url, host_room_url  # Whereby URLs
+    num_clients  # Updated via webhooks
+```
+
+## Jitsi Components
+
+### Core Architecture
+- **Jitsi Meet**: Web frontend (Next.js + React)
+- **Prosody**: XMPP server for messaging/rooms
+- **Jicofo**: Conference focus (orchestration)
+- **JVB**: Videobridge (media routing)
+- **Jibri**: Recording service
+- **Jigasi**: SIP gateway (optional, for phone dial-in)
+
+### Exposure Requirements
+- **Web service**: 443/80 (frontend)
+- **JVB**: 10000/UDP (media streams) - **MUST EXPOSE**
+- **Prosody**: 5280 (BOSH/WebSocket) - can proxy via web
+- **Jicofo, Jibri, Jigasi**: Internal only
+
+## Recording with Jibri
+
+### How Jibri Works
+- Each Jibri instance handles **one recording at a time**
+- Records mixed audio/video to MP4 format
+- Uses Chrome headless + ffmpeg for capture
+- Supports finalize scripts for post-processing
+
+### Jibri Pool for Scaling
+- Multiple Jibri instances join "jibribrewery" MUC
+- Jicofo distributes recording requests to available instances
+- Automatic load balancing and failover
+
+```yaml
+# Multiple Jibri instances
+jibri1:
+  environment:
+    - JIBRI_INSTANCE_ID=jibri1
+    - JIBRI_BREWERY_MUC=jibribrewery
+
+jibri2:
+  environment:
+    - JIBRI_INSTANCE_ID=jibri2
+    - JIBRI_BREWERY_MUC=jibribrewery
+```
+
+### Recording Automation Options
+1. **Environment Variables**: `ENABLE_RECORDING=1`, `AUTO_RECORDING=1`
+2. **URL Parameters**: `?config.autoRecord=true`
+3. **JWT Token**: Include recording permissions in JWT
+4. **API Control**: `api.executeCommand('startRecording')`
+
+### Post-Processing Integration
+```bash
+#!/bin/bash
+# finalize.sh - runs after recording completion
+RECORDING_FILE=$1
+MEETING_METADATA=$2
+ROOM_NAME=$3
+
+# Copy to Reflector-accessible location
+cp "$RECORDING_FILE" /shared/reflector-uploads/
+
+# Trigger Reflector processing
+curl -X POST "http://reflector-api:8000/v1/transcripts/process" \
+  -H "Content-Type: application/json" \
+  -d "{
+    \"file_path\": \"/shared/reflector-uploads/$(basename $RECORDING_FILE)\",
+    \"room_name\": \"$ROOM_NAME\",
+    \"source\": \"jitsi\"
+  }"
+```
+
+## React Integration
+
+### Official React SDK
+```bash
+npm i @jitsi/react-sdk
+```
+
+```jsx
+import { JitsiMeeting } from '@jitsi/react-sdk'
+
+<JitsiMeeting
+  room="meeting-room"
+  serverURL="https://your-jitsi.domain"
+  jwt="your-jwt-token"
+  config={{
+    startWithAudioMuted: true,
+    fileRecordingsEnabled: true,
+    autoRecord: true
+  }}
+  onParticipantJoined={(participant) => {
+    // Track participant events
+  }}
+  onRecordingStatusChanged={(status) => {
+    // Handle recording events
+  }}
+/>
+```
+
+## Authentication & Room Control
+
+### JWT-Based Access Control
+```python
+def generate_jitsi_jwt(payload):
+    return jwt.encode({
+        "aud": "jitsi",
+        "iss": "reflector",
+        "sub": "reflector-user",
+        "room": payload["room"],
+        "exp": int(payload["exp"].timestamp()),
+        "context": {
+            "user": {
+                "name": payload["user_name"],
+                "moderator": payload.get("moderator", False)
+            },
+            "features": {
+                "recording": payload.get("recording", True)
+            }
+        }
+    }, JITSI_JWT_SECRET)
+```
+
+### Prevent Anonymous Room Creation
+```bash
+# Environment configuration
+ENABLE_AUTH=1
+ENABLE_GUESTS=0
+AUTH_TYPE=jwt
+JWT_APP_ID=reflector
+JWT_APP_SECRET=your-secret-key
+```
+
+## Webhook Integration
+
+### Real-time Events via Prosody
+Custom event-sync module can send webhooks for:
+- Participant join/leave
+- Recording start/stop
+- Room creation/destruction
+- Mute/unmute events
+
+```lua
+-- mod_event_sync.lua
+module:hook("muc-occupant-joined", function(event)
+    send_event({
+        type = "participant_joined",
+        room = event.room.jid,
+        participant = {
+            nick = event.occupant.nick,
+            jid = event.occupant.jid,
+        },
+        timestamp = os.time(),
+    });
+end);
+```
+
+### Jibri Recording Webhooks
+```bash
+# Environment variable
+JIBRI_WEBHOOK_SUBSCRIBERS=https://your-reflector.com/webhooks/jibri
+```
+
+## Proposed Reflector Integration
+
+### Modified Database Schema
+```python
+class Meeting(BaseModel):
+    id: str  # Our generated meeting ID
+    room_name: str  # Generated: reflector-{room.name}-{timestamp}
+    room_url: str  # https://jitsi.domain/room_name?jwt=token
+    host_room_url: str  # Same but with moderator JWT
+    # Add Jitsi-specific fields
+    jitsi_jwt: str  # JWT token
+    jitsi_room_id: str  # Internal room identifier
+    recording_status: str  # pending, recording, completed
+    recording_file_path: Optional[str]
+```
+
+### API Replacement
+```python
+# Replace whereby.py with jitsi.py
+async def create_meeting(room_name_prefix: str, end_date: datetime, room: Room):
+    # Generate unique room name
+    jitsi_room = f"reflector-{room.name}-{int(time.time())}"
+
+    # Generate JWT tokens
+    user_jwt = generate_jwt(room=jitsi_room, moderator=False, exp=end_date)
+    host_jwt = generate_jwt(room=jitsi_room, moderator=True, exp=end_date)
+
+    return {
+        "meetingId": generate_uuid4(),  # Our ID
+        "roomName": jitsi_room,
+        "roomUrl": f"https://jitsi.domain/{jitsi_room}?jwt={user_jwt}",
+        "hostRoomUrl": f"https://jitsi.domain/{jitsi_room}?jwt={host_jwt}",
+        "startDate": datetime.now().isoformat(),
+        "endDate": end_date.isoformat(),
+    }
+```
+
+### Webhook Endpoints
+```python
+# Replace whereby webhook with jitsi webhooks
+@router.post("/jitsi/events")
+async def jitsi_events_webhook(event_data: dict):
+    event_type = event_data.get("event")
+    room_name = event_data.get("room", "").split("@")[0]
+
+    meeting = await Meeting.get_by_room(room_name)
+
+    if event_type == "muc-occupant-joined":
+        # Update participant count
+        meeting.num_clients += 1
+
+    elif event_type == "jibri-recording-on":
+        meeting.recording_status = "recording"
+
+    elif event_type == "jibri-recording-off":
+        meeting.recording_status = "processing"
+        await process_meeting_recording.delay(meeting.id)
+
+@router.post("/jibri/recording-complete")
+async def recording_complete(data: dict):
+    # Handle finalize script webhook
+    room_name = data.get("room_name")
+    file_path = data.get("file_path")
+
+    meeting = await Meeting.get_by_room(room_name)
+    meeting.recording_file_path = file_path
+    meeting.recording_status = "completed"
+
+    # Start Reflector processing
+    await process_recording_for_transcription(meeting.id, file_path)
+```
+
+## Deployment with Docker
+
+### Official docker-jitsi-meet
+```bash
+# Download official release
+wget $(wget -q -O - https://api.github.com/repos/jitsi/docker-jitsi-meet/releases/latest | grep zip | cut -d\" -f4)
+
+# Setup
+mkdir -p ~/.jitsi-meet-cfg/{web,transcripts,prosody/config,prosody/prosody-plugins-custom,jicofo,jvb,jigasi,jibri}
+./gen-passwords.sh  # Generate secure passwords
+docker compose up -d
+```
+
+### Coolify Integration
+```yaml
+services:
+  web:
+    ports: ["80:80", "443:443"]
+  jvb:
+    ports: ["10000:10000/udp"]  # Must expose for media
+  jibri1:
+    environment:
+      - JIBRI_INSTANCE_ID=jibri1
+      - JIBRI_FINALIZE_RECORDING_SCRIPT_PATH=/config/finalize.sh
+  jibri2:
+    environment:
+      - JIBRI_INSTANCE_ID=jibri2
+```
+
+## Benefits vs Whereby
+
+### Cost & Control
+✅ **No per-minute charges** - significant cost savings
+✅ **Full recording control** - direct file access
+✅ **Custom branding** - complete UI control
+✅ **Self-hosted** - no vendor lock-in
+
+### Technical Advantages
+✅ **Real-time events** - immediate webhook notifications
+✅ **Rich participant metadata** - detailed tracking
+✅ **JWT security** - token-based access with expiration
+✅ **Multiple recording formats** - audio-only options
+✅ **Scalable architecture** - horizontal Jibri scaling
+
+### Integration Benefits
+✅ **Same API surface** - minimal changes to existing code
+✅ **React SDK** - better frontend integration
+✅ **Direct processing** - no S3 download delays
+✅ **Event-driven architecture** - better real-time capabilities
+
+## Implementation Plan
+
+1. **Deploy Jitsi Stack** - Set up docker-jitsi-meet with multiple Jibri instances
+2. **Create jitsi.py** - Replace whereby.py with Jitsi API functions
+3. **Update Database** - Add Jitsi-specific fields to Meeting model
+4. **Webhook Integration** - Replace Whereby webhooks with Jitsi events
+5. **Frontend Updates** - Replace Whereby embed with Jitsi React SDK
+6. **Testing & Migration** - Gradual rollout with fallback to Whereby
+
+## Recording Limitations & Considerations
+
+### Current Limitations
+- **Mixed audio only** - Jibri doesn't separate participant tracks natively
+- **One recording per Jibri** - requires multiple instances for concurrent recordings
+- **Chrome dependency** - Jibri uses headless Chrome for recording
+
+### Metadata Capabilities
+✅ **Participant join/leave timestamps** - via webhooks
+✅ **Speaking time tracking** - via audio level events
+✅ **Meeting duration** - precise timing
+✅ **Room-specific data** - custom metadata in JWT
+
+### Alternative Recording Methods
+- **Local recording** - browser-based, per-participant
+- **Custom recording** - lib-jitsi-meet for individual streams
+- **Third-party solutions** - Recall.ai, Otter.ai integrations
+
+## Security Considerations
+
+### JWT Configuration
+- **Room-specific tokens** - limit access to specific rooms
+- **Time-based expiration** - automatic cleanup
+- **Feature permissions** - control recording, moderation rights
+- **User identification** - embed user metadata in tokens
+
+### Access Control
+- **No anonymous rooms** - all rooms require valid JWT
+- **API-only creation** - prevent direct room access
+- **Webhook verification** - HMAC signature validation
+
+## Next Steps
+
+1. **Deploy test Jitsi instance** - validate recording pipeline
+2. **Prototype jitsi.py** - create equivalent API functions
+3. **Test webhook integration** - ensure event delivery works
+4. **Performance testing** - validate multiple concurrent recordings
+5. **Migration strategy** - plan gradual transition from Whereby
+
+---
+
+*This document serves as the comprehensive planning and research notes for Jitsi integration in Reflector. It should be updated as implementation progresses and new insights are discovered.*

docs/video-jitsi.md

@@ -0,0 +1,720 @@
+# Jitsi Meet Integration Configuration Guide
+
+This guide explains how to configure Reflector to use your self-hosted Jitsi Meet installation for video meetings, recording, and participant tracking.
+
+## Overview
+
+Jitsi Meet is an open-source video conferencing platform that can be self-hosted. Reflector integrates with Jitsi Meet to:
+
+- Create secure meeting rooms with JWT authentication
+- Track participant join/leave events via Prosody webhooks
+- Record meetings using Jibri recording service
+- Process recordings for transcription and analysis
+
+## Requirements
+
+### Self-Hosted Jitsi Meet
+
+You need a complete Jitsi Meet installation including:
+
+1. **Jitsi Meet Web Interface** - The main meeting interface
+2. **Prosody XMPP Server** - Handles room management and authentication
+3. **Jicofo (JItsi COnference FOcus)** - Manages media sessions
+4. **Jitsi Videobridge (JVB)** - Handles WebRTC media routing
+5. **Jibri Recording Service** - Records meetings (optional but recommended)
+
+### System Requirements
+
+- **Domain with SSL Certificate** - Required for WebRTC functionality
+- **Prosody mod_event_sync** - For webhook event handling
+- **JWT Authentication** - For secure room access control
+- **Storage Solution** - For recording files (local or cloud)
+
+## Configuration Variables
+
+Add the following environment variables to your Reflector `.env` file:
+
+### Required Variables
+
+```bash
+# Jitsi Meet Domain (without https://)
+JITSI_DOMAIN=meet.example.com
+
+# JWT Secret for room authentication (generate with: openssl rand -hex 32)
+JITSI_JWT_SECRET=your-64-character-hex-secret-here
+
+# Webhook secret for event handling (generate with: openssl rand -hex 16)
+JITSI_WEBHOOK_SECRET=your-32-character-hex-secret-here
+```
+
+### Optional Variables
+
+```bash
+# Application identifier (should match Jitsi configuration)
+JITSI_APP_ID=reflector
+
+# JWT issuer and audience (should match Jitsi configuration)
+JITSI_JWT_ISSUER=reflector
+JITSI_JWT_AUDIENCE=jitsi
+```
+
+## Installation Steps
+
+### 1. Jitsi Meet Server Installation
+
+#### Quick Installation (Ubuntu/Debian)
+
+```bash
+# Add Jitsi repository
+curl -fsSL https://download.jitsi.org/jitsi-key.gpg.key | sudo gpg --dearmor -o /usr/share/keyrings/jitsi-keyring.gpg
+echo "deb [signed-by=/usr/share/keyrings/jitsi-keyring.gpg] https://download.jitsi.org stable/" | sudo tee /etc/apt/sources.list.d/jitsi-stable.list
+
+# Install Jitsi Meet
+sudo apt update
+sudo apt install jitsi-meet
+
+# Configure SSL certificate
+sudo /usr/share/jitsi-meet/scripts/install-letsencrypt-cert.sh
+```
+
+#### Docker Installation
+
+```bash
+# Clone Jitsi Docker repository
+git clone https://github.com/jitsi/docker-jitsi-meet
+cd docker-jitsi-meet
+
+# Copy environment template
+cp env.example .env
+
+# Edit configuration
+nano .env
+
+# Start services
+docker-compose up -d
+```
+
+### 2. JWT Authentication Setup
+
+#### Update Prosody Configuration
+
+Edit `/etc/prosody/conf.d/your-domain.cfg.lua`:
+
+```lua
+VirtualHost "meet.example.com"
+    authentication = "token"
+    app_id = "reflector"
+    app_secret = "your-jwt-secret-here"
+
+    -- Allow anonymous access for guests
+    c2s_require_encryption = false
+    admins = { "focusUser@auth.meet.example.com" }
+
+    modules_enabled = {
+        "bosh";
+        "pubsub";
+        "ping";
+        "roster";
+        "saslauth";
+        "tls";
+        "dialback";
+        "disco";
+        "carbons";
+        "pep";
+        "private";
+        "blocklist";
+        "vcard";
+        "version";
+        "uptime";
+        "time";
+        "ping";
+        "register";
+        "admin_adhoc";
+        "token_verification";
+        "event_sync";  -- Required for webhooks
+    }
+```
+
+#### Configure Jitsi Meet Interface
+
+Edit `/etc/jitsi/meet/your-domain-config.js`:
+
+```javascript
+var config = {
+    hosts: {
+        domain: 'meet.example.com',
+        muc: 'conference.meet.example.com'
+    },
+
+    // Enable JWT authentication
+    enableUserRolesBasedOnToken: true,
+
+    // Recording configuration
+    fileRecordingsEnabled: true,
+    liveStreamingEnabled: false,
+
+    // Reflector integration settings
+    prejoinPageEnabled: true,
+    requireDisplayName: true
+};
+```
+
+### 3. Webhook Event Configuration
+
+#### Install Event Sync Module
+
+```bash
+# Download the module
+cd /usr/share/jitsi-meet/prosody-plugins/
+wget https://raw.githubusercontent.com/jitsi-contrib/prosody-plugins/main/mod_event_sync.lua
+```
+
+#### Configure Event Sync
+
+Add to your Prosody configuration:
+
+```lua
+Component "conference.meet.example.com" "muc"
+    storage = "memory"
+    modules_enabled = {
+        "muc_meeting_id";
+        "muc_domain_mapper";
+        "polls";
+        "event_sync";  -- Enable event sync
+    }
+
+    -- Event sync webhook configuration
+    event_sync_url = "https://your-reflector-domain.com/v1/jitsi/events"
+    event_sync_secret = "your-webhook-secret-here"
+
+    -- Events to track
+    event_sync_events = {
+        "muc-occupant-joined",
+        "muc-occupant-left",
+        "jibri-recording-on",
+        "jibri-recording-off"
+    }
+
+#### Webhook Event Payload Examples
+
+**Participant Joined Event:**
+```json
+{
+  "event": "muc-occupant-joined",
+  "room": "reflector-my-room-uuid123",
+  "timestamp": "2025-01-15T10:30:00.000Z",
+  "data": {
+    "occupant_id": "participant-456",
+    "nick": "John Doe",
+    "role": "participant",
+    "affiliation": "none"
+  }
+}
+```
+
+**Recording Started Event:**
+```json
+{
+  "event": "jibri-recording-on",
+  "room": "reflector-my-room-uuid123",
+  "timestamp": "2025-01-15T10:32:00.000Z",
+  "data": {
+    "recording_id": "rec-789",
+    "initiator": "moderator-123"
+  }
+}
+```
+
+**Recording Completed Event:**
+```json
+{
+  "room_name": "reflector-my-room-uuid123",
+  "recording_file": "/var/recordings/rec-789.mp4",
+  "recording_status": "completed",
+  "timestamp": "2025-01-15T11:15:00.000Z"
+}
+```
+
+### 4. Jibri Recording Setup (Optional)
+
+#### Install Jibri
+
+```bash
+# Install Jibri package
+sudo apt install jibri
+
+# Create recording directory
+sudo mkdir -p /var/recordings
+sudo chown jibri:jibri /var/recordings
+```
+
+#### Configure Jibri
+
+Edit `/etc/jitsi/jibri/jibri.conf`:
+
+```hocon
+jibri {
+    recording {
+        recordings-directory = "/var/recordings"
+        finalize-script = "/opt/jitsi/jibri/finalize.sh"
+    }
+
+    api {
+        xmpp {
+            environments = [{
+                name = "prod environment"
+                xmpp-server-hosts = ["meet.example.com"]
+                xmpp-domain = "meet.example.com"
+
+                control-muc {
+                    domain = "internal.auth.meet.example.com"
+                    room-name = "JibriBrewery"
+                    nickname = "jibri-nickname"
+                }
+
+                control-login {
+                    domain = "auth.meet.example.com"
+                    username = "jibri"
+                    password = "jibri-password"
+                }
+            }]
+        }
+    }
+}
+```
+
+#### Create Finalize Script
+
+Create `/opt/jitsi/jibri/finalize.sh`:
+
+```bash
+#!/bin/bash
+# Jibri finalize script for Reflector integration
+
+RECORDING_FILE="$1"
+ROOM_NAME="$2"
+REFLECTOR_API_URL="${REFLECTOR_API_URL:-http://localhost:1250}"
+
+# Prepare webhook payload
+TIMESTAMP=$(date -u +%Y-%m-%dT%H:%M:%S.%3NZ)
+PAYLOAD=$(cat <<EOF
+{
+    "room_name": "$ROOM_NAME",
+    "recording_file": "$RECORDING_FILE",
+    "recording_status": "completed",
+    "timestamp": "$TIMESTAMP"
+}
+EOF
+)
+
+# Generate signature
+SIGNATURE=$(echo -n "$PAYLOAD" | openssl dgst -sha256 -hmac "$JITSI_WEBHOOK_SECRET" | cut -d' ' -f2)
+
+# Send webhook to Reflector
+curl -X POST "$REFLECTOR_API_URL/v1/jibri/recording-complete" \
+    -H "Content-Type: application/json" \
+    -H "X-Jitsi-Signature: $SIGNATURE" \
+    -d "$PAYLOAD"
+
+echo "Recording finalization webhook sent for room: $ROOM_NAME"
+```
+
+Make executable:
+```bash
+sudo chmod +x /opt/jitsi/jibri/finalize.sh
+```
+
+### 5. Restart Services
+
+After configuration changes:
+
+```bash
+sudo systemctl restart prosody
+sudo systemctl restart jicofo
+sudo systemctl restart jitsi-videobridge2
+sudo systemctl restart jibri
+sudo systemctl restart nginx
+```
+
+## Room Configuration
+
+### Creating Jitsi Rooms
+
+Create rooms with Jitsi platform in Reflector:
+
+```bash
+curl -X POST "https://your-reflector-domain.com/v1/rooms" \
+  -H "Content-Type: application/json" \
+  -H "Authorization: Bearer $AUTH_TOKEN" \
+  -d '{
+    "name": "my-jitsi-room",
+    "platform": "jitsi",
+    "recording_type": "cloud",
+    "recording_trigger": "automatic-2nd-participant",
+    "is_locked": false,
+    "room_mode": "normal"
+  }'
+```
+
+### Meeting Creation
+
+Meetings automatically use JWT authentication:
+
+```bash
+curl -X POST "https://your-reflector-domain.com/v1/rooms/my-jitsi-room/meeting" \
+  -H "Authorization: Bearer $AUTH_TOKEN"
+```
+
+Response includes JWT-authenticated URLs:
+```json
+{
+  "id": "meeting-uuid",
+  "room_name": "reflector-my-jitsi-room-123456",
+  "room_url": "https://meet.example.com/room?jwt=user-token",
+  "host_room_url": "https://meet.example.com/room?jwt=moderator-token"
+}
+```
+
+## Features and Capabilities
+
+### JWT Authentication
+
+Reflector automatically generates JWT tokens with:
+- **Room Access Control** - Secure room entry
+- **User Roles** - Moderator vs participant permissions
+- **Expiration** - Configurable token lifetime (default 8 hours)
+- **Custom Claims** - Room-specific metadata
+
+### Recording Options
+
+**Recording Types:**
+- `"none"` - No recording
+- `"local"` - Local Jibri recording
+- `"cloud"` - Cloud recording (requires external storage)
+
+**Recording Triggers:**
+- `"none"` - Manual recording only
+- `"prompt"` - Prompt users to start
+- `"automatic"` - Start immediately
+- `"automatic-2nd-participant"` - Start when 2nd person joins
+
+### Event Tracking and Storage
+
+Reflector automatically stores all webhook events in the `meetings` table for comprehensive meeting analytics:
+
+**Supported Event Types:**
+- `muc-occupant-joined` - Participant joined the meeting
+- `muc-occupant-left` - Participant left the meeting
+- `jibri-recording-on` - Recording started
+- `jibri-recording-off` - Recording stopped
+- `recording_completed` - Recording file ready for processing
+
+**Event Storage Structure:**
+Each webhook event is stored as a JSON object in the `meetings.events` column:
+```json
+{
+  "type": "muc-occupant-joined",
+  "timestamp": "2025-01-15T10:30:00.123456Z",
+  "data": {
+    "timestamp": "2025-01-15T10:30:00Z",
+    "user_id": "participant-123",
+    "display_name": "John Doe"
+  }
+}
+```
+
+**Querying Stored Events:**
+```sql
+-- Get all events for a meeting
+SELECT events FROM meeting WHERE id = 'meeting-uuid';
+
+-- Count participant joins
+SELECT json_array_length(
+  json_extract(events, '$[*] ? (@.type == "muc-occupant-joined")')
+) as total_joins FROM meeting WHERE id = 'meeting-uuid';
+```
+
+## Testing and Verification
+
+### Health Check
+
+Test Jitsi webhook integration:
+
+```bash
+curl "https://your-reflector-domain.com/v1/jitsi/health"
+```
+
+Expected response:
+```json
+{
+    "status": "ok",
+    "service": "jitsi-webhooks",
+    "timestamp": "2025-01-15T10:30:00.000Z",
+    "webhook_secret_configured": true
+}
+```
+
+### JWT Token Testing
+
+Verify JWT generation works:
+```bash
+# Create a test meeting
+MEETING=$(curl -X POST "https://your-reflector-domain.com/v1/rooms/test-room/meeting" \
+  -H "Authorization: Bearer $AUTH_TOKEN" | jq -r '.room_url')
+
+echo "Test meeting URL: $MEETING"
+```
+
+### Webhook Testing
+
+#### Manual Webhook Event Testing
+
+Test participant join event:
+```bash
+# Generate proper signature
+PAYLOAD='{"event":"muc-occupant-joined","room":"reflector-test-room-uuid","timestamp":"2025-01-15T10:30:00.000Z","data":{"user_id":"test-user","display_name":"Test User"}}'
+SIGNATURE=$(echo -n "$PAYLOAD" | openssl dgst -sha256 -hmac "$JITSI_WEBHOOK_SECRET" | cut -d' ' -f2)
+
+curl -X POST "https://your-reflector-domain.com/v1/jitsi/events" \
+  -H "Content-Type: application/json" \
+  -H "X-Jitsi-Signature: $SIGNATURE" \
+  -d "$PAYLOAD"
+```
+
+Expected response:
+```json
+{
+  "status": "ok",
+  "event": "muc-occupant-joined",
+  "room": "reflector-test-room-uuid"
+}
+```
+
+#### Recording Webhook Testing
+
+Test recording completion event:
+```bash
+PAYLOAD='{"room_name":"reflector-test-room-uuid","recording_file":"/recordings/test.mp4","recording_status":"completed","timestamp":"2025-01-15T10:30:00.000Z"}'
+SIGNATURE=$(echo -n "$PAYLOAD" | openssl dgst -sha256 -hmac "$JITSI_WEBHOOK_SECRET" | cut -d' ' -f2)
+
+curl -X POST "https://your-reflector-domain.com/v1/jibri/recording-complete" \
+  -H "Content-Type: application/json" \
+  -H "X-Jitsi-Signature: $SIGNATURE" \
+  -d "$PAYLOAD"
+```
+
+#### Event Storage Verification
+
+Verify events were stored:
+```bash
+# Check meeting events via API (requires authentication)
+curl -H "Authorization: Bearer $AUTH_TOKEN" \
+  "https://your-reflector-domain.com/v1/meetings/{meeting-id}"
+```
+
+## Troubleshooting
+
+### Common Issues
+
+#### JWT Authentication Failures
+
+**Symptoms**: Users cannot join rooms, "Authentication failed" errors
+
+**Solutions**:
+1. Verify `JITSI_JWT_SECRET` matches Prosody configuration
+2. Check JWT token hasn't expired (default 8 hours)
+3. Ensure system clocks are synchronized between servers
+4. Validate JWT issuer/audience configuration matches
+
+**Debug JWT tokens**:
+```bash
+# Decode JWT payload
+echo "JWT_TOKEN_HERE" | cut -d'.' -f2 | base64 -d | jq
+```
+
+#### Webhook Events Not Received
+
+**Symptoms**: Participant counts not updating, no recording events
+
+**Solutions**:
+1. Verify `mod_event_sync` is loaded in Prosody
+2. Check webhook URL is accessible from Jitsi server
+3. Validate webhook signature generation
+4. Review Prosody and Reflector logs
+
+**Debug webhook connectivity**:
+```bash
+# Test from Jitsi server
+curl -v "https://your-reflector-domain.com/v1/jitsi/health"
+
+# Check Prosody logs
+sudo tail -f /var/log/prosody/prosody.log
+```
+
+#### Webhook Signature Verification Issues
+
+**Symptoms**: HTTP 401 "Invalid webhook signature" errors
+
+**Solutions**:
+1. Verify webhook secret matches between Jitsi and Reflector
+2. Check payload encoding (no extra whitespace)
+3. Ensure proper HMAC-SHA256 signature generation
+
+**Debug signature generation**:
+```bash
+# Test signature manually
+PAYLOAD='{"event":"test","room":"test","timestamp":"2025-01-15T10:30:00.000Z","data":{}}'
+SECRET="your-webhook-secret-here"
+
+# Generate signature (should match X-Jitsi-Signature header)
+echo -n "$PAYLOAD" | openssl dgst -sha256 -hmac "$SECRET" | cut -d' ' -f2
+
+# Test with curl
+curl -X POST "https://your-reflector-domain.com/v1/jitsi/events" \
+  -H "Content-Type: application/json" \
+  -H "X-Jitsi-Signature: $(echo -n "$PAYLOAD" | openssl dgst -sha256 -hmac "$SECRET" | cut -d' ' -f2)" \
+  -d "$PAYLOAD" -v
+```
+
+#### Event Storage Problems
+
+**Symptoms**: Events received but not stored in database
+
+**Solutions**:
+1. Check database connectivity and permissions
+2. Verify meeting exists before event processing
+3. Review Reflector application logs
+4. Ensure JSON column support in database
+
+**Debug event storage**:
+```bash
+# Check meeting exists
+curl -H "Authorization: Bearer $TOKEN" \
+  "https://your-reflector-domain.com/v1/meetings/{meeting-id}"
+
+# Monitor database queries (if using PostgreSQL)
+sudo -u postgres psql -c "SELECT * FROM pg_stat_activity WHERE query LIKE '%meeting%';"
+
+# Check Reflector logs for event processing
+sudo journalctl -u reflector -f | grep -E "(event|webhook|jitsi)"
+```
+
+#### Recording Issues
+
+**Symptoms**: Recordings not starting, finalize script errors
+
+**Solutions**:
+1. Verify Jibri service status: `sudo systemctl status jibri`
+2. Check recording directory permissions: `/var/recordings`
+3. Validate finalize script execution permissions
+4. Monitor Jibri logs: `sudo journalctl -u jibri -f`
+
+**Test finalize script**:
+```bash
+sudo -u jibri /opt/jitsi/jibri/finalize.sh "/test/recording.mp4" "test-room"
+```
+
+#### Meeting Creation Failures
+
+**Symptoms**: HTTP 500 errors when creating meetings
+
+**Solutions**:
+1. Check Reflector logs for JWT generation errors
+2. Verify all required environment variables are set
+3. Ensure Jitsi domain is accessible from Reflector
+4. Test JWT secret configuration
+
+### Debug Commands
+
+```bash
+# Verify Prosody configuration
+sudo prosodyctl check config
+
+# Check Jitsi services status
+sudo systemctl status prosody jicofo jitsi-videobridge2
+
+# Test JWT generation
+curl -X POST "https://your-reflector-domain.com/v1/rooms/test/meeting" \
+  -H "Authorization: Bearer $TOKEN" -v
+
+# Monitor webhook events
+sudo tail -f /var/log/reflector/app.log | grep jitsi
+
+# Check SSL certificates
+sudo certbot certificates
+```
+
+### Performance Optimization
+
+#### Scaling Considerations
+
+**Single Server Limits:**
+- ~50 concurrent participants per JVB instance
+- ~10 concurrent Jibri recordings
+- CPU and bandwidth become bottlenecks
+
+**Multi-Server Setup:**
+- Multiple JVB instances for scaling
+- Dedicated Jibri recording servers
+- Load balancing for high availability
+
+#### Resource Monitoring
+
+```bash
+# Monitor JVB performance
+sudo systemctl status jitsi-videobridge2
+sudo journalctl -u jitsi-videobridge2 -f
+
+# Check Prosody connections
+sudo prosodyctl mod_admin_telnet
+> c2s:show()
+> muc:rooms()
+```
+
+## Security Best Practices
+
+### JWT Security
+- Use strong, unique secrets (32+ characters)
+- Rotate JWT secrets regularly
+- Implement proper token expiration
+- Never log or expose JWT tokens
+
+### Network Security
+- Use HTTPS/WSS for all communications
+- Implement proper firewall rules
+- Consider VPN for server-to-server communication
+- Monitor for unauthorized access attempts
+
+### Recording Security
+- Encrypt recordings at rest
+- Implement access controls for recording files
+- Regular security audits of file permissions
+- Comply with data protection regulations
+
+## Migration from Whereby
+
+If migrating from Whereby to Jitsi:
+
+1. **Parallel Setup** - Configure Jitsi alongside existing Whereby
+2. **Room Migration** - Update room platform field to "jitsi"
+3. **Test Integration** - Verify meeting creation and webhooks
+4. **User Training** - Different UI and feature set
+5. **Monitor Performance** - Watch for issues during transition
+6. **Cleanup** - Remove Whereby configuration when stable
+
+## Support and Resources
+
+### Jitsi Community Resources
+- **Documentation**: [jitsi.github.io/handbook](https://jitsi.github.io/handbook/)
+- **Community Forum**: [community.jitsi.org](https://community.jitsi.org/)
+- **GitHub Issues**: [github.com/jitsi/jitsi-meet](https://github.com/jitsi/jitsi-meet)
+
+### Professional Support
+- **8x8 Commercial Support** - Professional Jitsi hosting and support
+- **Community Consulting** - Third-party Jitsi implementation services
+
+### Monitoring and Maintenance
+- Monitor system resources (CPU, memory, bandwidth)
+- Regular security updates for all components
+- Backup configuration files and certificates
+- Test disaster recovery procedures

docs/video-whereby.md

@@ -0,0 +1,276 @@
+# Whereby Integration Configuration Guide
+
+This guide explains how to configure Reflector to use Whereby as your video meeting platform for room creation, recording, and participant tracking.
+
+## Overview
+
+Whereby is a browser-based video meeting platform that provides hosted meeting rooms with recording capabilities. Reflector integrates with Whereby's API to:
+
+- Create secure meeting rooms with custom branding
+- Handle participant join/leave events via webhooks
+- Automatically record meetings to AWS S3 storage
+- Track meeting sessions and participant counts
+
+## Requirements
+
+### Whereby Account Setup
+
+1. **Whereby Account**: Sign up for a Whereby business account at [whereby.com](https://whereby.com/business)
+2. **API Access**: Request API access from Whereby support (required for programmatic room creation)
+3. **Webhook Configuration**: Configure webhooks in your Whereby dashboard to point to your Reflector instance
+
+### AWS S3 Storage
+
+Whereby requires AWS S3 for recording storage. You need:
+- AWS account with S3 access
+- Dedicated S3 bucket for Whereby recordings
+- AWS IAM credentials with S3 write permissions
+
+## Configuration Variables
+
+Add the following environment variables to your Reflector `.env` file:
+
+### Required Variables
+
+```bash
+# Whereby API Configuration
+WHEREBY_API_KEY=your-whereby-jwt-api-key
+WHEREBY_WEBHOOK_SECRET=your-webhook-secret-from-whereby
+
+# AWS S3 Storage for Recordings
+AWS_WHEREBY_ACCESS_KEY_ID=your-aws-access-key
+AWS_WHEREBY_ACCESS_KEY_SECRET=your-aws-secret-key
+RECORDING_STORAGE_AWS_BUCKET_NAME=your-s3-bucket-name
+```
+
+### Optional Variables
+
+```bash
+# Whereby API URL (defaults to production)
+WHEREBY_API_URL=https://api.whereby.dev/v1
+
+# SQS Configuration (for recording processing)
+AWS_PROCESS_RECORDING_QUEUE_URL=https://sqs.region.amazonaws.com/account/queue
+SQS_POLLING_TIMEOUT_SECONDS=60
+```
+
+## Configuration Steps
+
+### 1. Whereby API Key Setup
+
+1. **Contact Whereby Support** to request API access for your account
+2. **Generate JWT Token** in your Whereby dashboard under API settings
+3. **Copy the JWT token** and set it as `WHEREBY_API_KEY` in your environment
+
+The API key is a JWT token that looks like:
+```
+eyJ[...truncated JWT token...]
+```
+
+### 2. Webhook Configuration
+
+1. **Access Whereby Dashboard** and navigate to webhook settings
+2. **Set Webhook URL** to your Reflector instance:
+   ```
+   https://your-reflector-domain.com/v1/whereby
+   ```
+3. **Configure Events** to send the following event types:
+   - `room.client.joined` - When participants join
+   - `room.client.left` - When participants leave
+4. **Generate Webhook Secret** and set it as `WHEREBY_WEBHOOK_SECRET`
+5. **Save Configuration** in your Whereby dashboard
+
+### 3. AWS S3 Storage Setup
+
+1. **Create S3 Bucket** dedicated for Whereby recordings
+2. **Create IAM User** with programmatic access
+3. **Attach S3 Policy** with the following permissions:
+   ```json
+   {
+     "Version": "2012-10-17",
+     "Statement": [
+       {
+         "Effect": "Allow",
+         "Action": [
+           "s3:PutObject",
+           "s3:PutObjectAcl",
+           "s3:GetObject"
+         ],
+         "Resource": "arn:aws:s3:::your-bucket-name/*"
+       }
+     ]
+   }
+   ```
+4. **Configure Environment Variables** with the IAM credentials
+
+### 4. Room Configuration
+
+When creating rooms in Reflector, set the platform to use Whereby:
+
+```bash
+curl -X POST "https://your-reflector-domain.com/v1/rooms" \
+  -H "Content-Type: application/json" \
+  -H "Authorization: Bearer $AUTH_TOKEN" \
+  -d '{
+    "name": "my-whereby-room",
+    "platform": "whereby",
+    "recording_type": "cloud",
+    "recording_trigger": "automatic-2nd-participant",
+    "is_locked": false,
+    "room_mode": "normal"
+  }'
+```
+
+## Meeting Features
+
+### Recording Options
+
+Whereby supports three recording types:
+- **`none`**: No recording
+- **`local`**: Local recording (not recommended for production)
+- **`cloud`**: Cloud recording to S3 (recommended)
+
+### Recording Triggers
+
+Control when recordings start:
+- **`none`**: No automatic recording
+- **`prompt`**: Prompt users to start recording
+- **`automatic`**: Start immediately when meeting begins
+- **`automatic-2nd-participant`**: Start when second participant joins
+
+### Room Modes
+
+- **`normal`**: Standard meeting room
+- **`group`**: Group meeting with advanced features
+
+## Webhook Event Handling
+
+Reflector automatically handles these Whereby webhook events:
+
+### Participant Tracking
+```json
+{
+  "type": "room.client.joined",
+  "data": {
+    "meetingId": "room-uuid",
+    "numClients": 2
+  }
+}
+```
+
+### Recording Events
+Whereby sends recording completion events that trigger Reflector's processing pipeline:
+- Audio transcription
+- Speaker diarization
+- Summary generation
+
+## Troubleshooting
+
+### Common Issues
+
+#### API Authentication Errors
+**Symptoms**: 401 Unauthorized errors when creating meetings
+
+**Solutions**:
+1. Verify your `WHEREBY_API_KEY` is correct and not expired
+2. Ensure you have API access enabled on your Whereby account
+3. Contact Whereby support if API access is not available
+
+#### Webhook Signature Validation Failed
+**Symptoms**: Webhook events rejected with 401 errors
+
+**Solutions**:
+1. Verify `WHEREBY_WEBHOOK_SECRET` matches your Whereby dashboard configuration
+2. Check webhook URL is correctly configured in Whereby dashboard
+3. Ensure webhook endpoint is accessible from Whereby servers
+
+#### Recording Upload Failures
+**Symptoms**: Recordings not appearing in S3 bucket
+
+**Solutions**:
+1. Verify AWS credentials have S3 write permissions
+2. Check S3 bucket name is correct and accessible
+3. Ensure AWS region settings match your bucket location
+4. Review AWS CloudTrail logs for permission issues
+
+#### Participant Count Not Updating
+**Symptoms**: Meeting participant counts remain at 0
+
+**Solutions**:
+1. Verify webhook events are being received at `/v1/whereby`
+2. Check webhook signature validation is passing
+3. Ensure meeting IDs match between Whereby and Reflector database
+
+### Debug Commands
+
+```bash
+# Test Whereby API connectivity
+curl -H "Authorization: Bearer $WHEREBY_API_KEY" \
+  https://api.whereby.dev/v1/meetings
+
+# Check webhook endpoint health
+curl https://your-reflector-domain.com/v1/whereby/health
+
+# Verify S3 bucket access
+aws s3 ls s3://your-bucket-name --profile whereby-user
+```
+
+## Security Considerations
+
+### API Key Security
+- Store API keys securely using environment variables
+- Rotate API keys regularly
+- Never commit API keys to version control
+- Use separate keys for development and production
+
+### Webhook Security
+- Always validate webhook signatures using HMAC-SHA256
+- Use HTTPS for all webhook endpoints
+- Implement rate limiting on webhook endpoints
+- Monitor webhook events for suspicious activity
+
+### Recording Privacy
+- Ensure S3 bucket access is restricted to authorized users
+- Consider encryption at rest for sensitive recordings
+- Implement retention policies for recorded content
+- Comply with data protection regulations (GDPR, etc.)
+
+## Performance Optimization
+
+### Meeting Scaling
+- Monitor concurrent meeting limits on your Whereby plan
+- Implement meeting cleanup for expired sessions
+- Use appropriate room modes for different use cases
+
+### Recording Processing
+- Configure SQS for asynchronous recording processing
+- Monitor S3 storage usage and costs
+- Implement automatic cleanup of processed recordings
+
+### Webhook Reliability
+- Implement webhook retry mechanisms
+- Monitor webhook delivery success rates
+- Log webhook events for debugging and auditing
+
+## Migration from Other Platforms
+
+If migrating from another video platform:
+
+1. **Update Room Configuration**: Change existing rooms to use `"platform": "whereby"`
+2. **Configure Webhooks**: Set up Whereby webhook endpoints
+3. **Test Integration**: Verify meeting creation and event handling
+4. **Monitor Performance**: Watch for any issues during transition
+5. **Update Documentation**: Inform users of any workflow changes
+
+## Support
+
+For Whereby-specific issues:
+- **Whereby Support**: [whereby.com/support](https://whereby.com/support)
+- **API Documentation**: [whereby.dev](https://whereby.dev)
+- **Status Page**: [status.whereby.com](https://status.whereby.com)
+
+For Reflector integration issues:
+- Check application logs for error details
+- Verify environment variable configuration
+- Test webhook connectivity and authentication
+- Review AWS permissions and S3 access

docs/video_platforms.md

@@ -0,0 +1,474 @@
+# Video Platforms Architecture (PR #529 Analysis)
+
+This document analyzes the video platforms refactoring implemented in PR #529 for daily.co integration, providing a blueprint for extending support to Jitsi and other video conferencing platforms.
+
+## Overview
+
+The video platforms refactoring introduces a clean abstraction layer that allows Reflector to support multiple video conferencing providers (Whereby, Daily.co, etc.) without changing core application logic. This architecture enables:
+
+- Seamless switching between video platforms
+- Platform-specific feature support
+- Isolated platform code organization
+- Consistent API surface across platforms
+- Feature flags for gradual migration
+
+## Architecture Components
+
+### 1. **Directory Structure**
+
+```
+server/reflector/video_platforms/
+├── __init__.py              # Public API exports
+├── base.py                  # Abstract base classes
+├── factory.py               # Platform client factory
+├── registry.py              # Platform registration system
+├── whereby.py               # Whereby implementation
+├── daily.py                 # Daily.co implementation
+└── mock.py                  # Testing implementation
+```
+
+### 2. **Core Abstract Classes**
+
+#### `VideoPlatformClient` (base.py)
+Abstract base class defining the interface all platforms must implement:
+
+```python
+class VideoPlatformClient(ABC):
+    PLATFORM_NAME: str = ""
+
+    @abstractmethod
+    async def create_meeting(self, room_name_prefix: str, end_date: datetime, room: Room) -> MeetingData
+
+    @abstractmethod
+    async def get_room_sessions(self, room_name: str) -> Dict[str, Any]
+
+    @abstractmethod
+    async def delete_room(self, room_name: str) -> bool
+
+    @abstractmethod
+    async def upload_logo(self, room_name: str, logo_path: str) -> bool
+
+    @abstractmethod
+    def verify_webhook_signature(self, body: bytes, signature: str, timestamp: Optional[str] = None) -> bool
+```
+
+#### `MeetingData` (base.py)
+Standardized meeting data structure returned by all platforms:
+
+```python
+class MeetingData(BaseModel):
+    meeting_id: str
+    room_name: str
+    room_url: str
+    host_room_url: str
+    platform: str
+    extra_data: Dict[str, Any] = {}  # Platform-specific data
+```
+
+#### `VideoPlatformConfig` (base.py)
+Unified configuration structure for all platforms:
+
+```python
+class VideoPlatformConfig(BaseModel):
+    api_key: str
+    webhook_secret: str
+    api_url: Optional[str] = None
+    subdomain: Optional[str] = None
+    s3_bucket: Optional[str] = None
+    s3_region: Optional[str] = None
+    aws_role_arn: Optional[str] = None
+    aws_access_key_id: Optional[str] = None
+    aws_access_key_secret: Optional[str] = None
+```
+
+### 3. **Platform Registration System**
+
+#### Registry Pattern (registry.py)
+- Automatic registration of built-in platforms
+- Runtime platform discovery
+- Type-safe client instantiation
+
+```python
+# Auto-registration of platforms
+_PLATFORMS: Dict[str, Type[VideoPlatformClient]] = {}
+
+def register_platform(name: str, client_class: Type[VideoPlatformClient])
+def get_platform_client(platform: str, config: VideoPlatformConfig) -> VideoPlatformClient
+```
+
+#### Factory System (factory.py)
+- Configuration management per platform
+- Platform selection logic
+- Feature flag integration
+
+```python
+def get_platform_for_room(room_id: Optional[str] = None) -> str:
+    """Determine which platform to use based on feature flags."""
+    if not settings.DAILY_MIGRATION_ENABLED:
+        return "whereby"
+
+    if room_id and room_id in settings.DAILY_MIGRATION_ROOM_IDS:
+        return "daily"
+
+    return settings.DEFAULT_VIDEO_PLATFORM
+```
+
+### 4. **Database Schema Changes**
+
+#### Room Model Updates
+Added `platform` field to track which video platform each room uses:
+
+```python
+# Database Schema
+platform_column = sqlalchemy.Column(
+    "platform",
+    sqlalchemy.String,
+    nullable=False,
+    server_default="whereby"
+)
+
+# Pydantic Model
+class Room(BaseModel):
+    platform: Literal["whereby", "daily"] = "whereby"
+```
+
+#### Meeting Model Updates
+Added `platform` field to meetings for tracking and debugging:
+
+```python
+# Database Schema
+platform_column = sqlalchemy.Column(
+    "platform",
+    sqlalchemy.String,
+    nullable=False,
+    server_default="whereby"
+)
+
+# Pydantic Model
+class Meeting(BaseModel):
+    platform: Literal["whereby", "daily"] = "whereby"
+```
+
+**Key Decision**: No platform-specific fields were added to models. Instead, the `extra_data` field in `MeetingData` handles platform-specific information, following the user's rule of using generic `provider_data` as JSON if needed.
+
+### 5. **Settings Configuration**
+
+#### Feature Flags
+```python
+# Migration control
+DAILY_MIGRATION_ENABLED: bool = True
+DAILY_MIGRATION_ROOM_IDS: list[str] = []
+DEFAULT_VIDEO_PLATFORM: str = "daily"
+
+# Daily.co specific settings
+DAILY_API_KEY: str | None = None
+DAILY_WEBHOOK_SECRET: str | None = None
+DAILY_SUBDOMAIN: str | None = None
+AWS_DAILY_S3_BUCKET: str | None = None
+AWS_DAILY_S3_REGION: str = "us-west-2"
+AWS_DAILY_ROLE_ARN: str | None = None
+```
+
+#### Configuration Pattern
+Each platform gets its own configuration namespace while sharing common patterns:
+
+```python
+def get_platform_config(platform: str) -> VideoPlatformConfig:
+    if platform == "whereby":
+        return VideoPlatformConfig(
+            api_key=settings.WHEREBY_API_KEY or "",
+            webhook_secret=settings.WHEREBY_WEBHOOK_SECRET or "",
+            # ... whereby-specific config
+        )
+    elif platform == "daily":
+        return VideoPlatformConfig(
+            api_key=settings.DAILY_API_KEY or "",
+            webhook_secret=settings.DAILY_WEBHOOK_SECRET or "",
+            # ... daily-specific config
+        )
+```
+
+### 6. **API Integration Updates**
+
+#### Room Creation (views/rooms.py)
+Updated to use platform factory instead of direct Whereby calls:
+
+```python
+@router.post("/rooms/{room_name}/meeting")
+async def rooms_create_meeting(room_name: str, user: UserInfo):
+    # OLD: Direct Whereby integration
+    # whereby_meeting = await create_meeting("", end_date=end_date, room=room)
+
+    # NEW: Platform abstraction
+    platform = get_platform_for_room(room.id)
+    client = create_platform_client(platform)
+
+    meeting_data = await client.create_meeting(
+        room_name_prefix=room.name, end_date=end_date, room=room
+    )
+
+    await client.upload_logo(meeting_data.room_name, "./images/logo.png")
+```
+
+### 7. **Webhook Handling**
+
+#### Separate Webhook Endpoints
+Each platform gets its own webhook endpoint with platform-specific signature verification:
+
+```python
+# views/daily.py
+@router.post("/daily_webhook")
+async def daily_webhook(event: DailyWebhookEvent, request: Request):
+    # Verify Daily.co signature
+    body = await request.body()
+    signature = request.headers.get("X-Daily-Signature", "")
+
+    if not verify_daily_webhook_signature(body, signature):
+        raise HTTPException(status_code=401)
+
+    # Handle platform-specific events
+    if event.type == "participant.joined":
+        await _handle_participant_joined(event)
+```
+
+#### Consistent Event Handling
+Despite different event formats, the core business logic remains the same:
+
+```python
+async def _handle_participant_joined(event):
+    room_name = event.data.get("room", {}).get("name")  # Daily.co format
+    meeting = await meetings_controller.get_by_room_name(room_name)
+    if meeting:
+        current_count = getattr(meeting, "num_clients", 0)
+        await meetings_controller.update_meeting(
+            meeting.id, num_clients=current_count + 1
+        )
+```
+
+### 8. **Worker Task Integration**
+
+#### New Task for Daily.co Recording Processing
+Added platform-specific recording processing while maintaining the same pipeline:
+
+```python
+@shared_task
+@asynctask
+async def process_recording_from_url(recording_url: str, meeting_id: str, recording_id: str):
+    """Process recording from Direct URL (Daily.co webhook)."""
+    logger.info("Processing recording from URL for meeting: %s", meeting_id)
+    # Uses same processing pipeline as Whereby S3 recordings
+```
+
+**Key Decision**: Worker tasks remain in main worker module but could be moved to platform-specific folders as suggested by the user.
+
+### 9. **Testing Infrastructure**
+
+#### Comprehensive Test Suite
+- Unit tests for each platform client
+- Integration tests for platform switching
+- Mock platform for testing without external dependencies
+- Webhook signature verification tests
+
+```python
+class TestPlatformIntegration:
+    """Integration tests for platform switching."""
+
+    async def test_platform_switching_preserves_interface(self):
+        """Test that different platforms provide consistent interface."""
+        # Test both Mock and Daily platforms return MeetingData objects
+        # with consistent fields
+```
+
+## Implementation Patterns for Jitsi Integration
+
+Based on the daily.co implementation, here's how Jitsi should be integrated:
+
+### 1. **Jitsi Client Implementation**
+
+```python
+# video_platforms/jitsi.py
+class JitsiClient(VideoPlatformClient):
+    PLATFORM_NAME = "jitsi"
+
+    async def create_meeting(self, room_name_prefix: str, end_date: datetime, room: Room) -> MeetingData:
+        # Generate unique room name
+        jitsi_room = f"reflector-{room.name}-{int(time.time())}"
+
+        # Generate JWT tokens
+        user_jwt = self._generate_jwt(room=jitsi_room, moderator=False, exp=end_date)
+        host_jwt = self._generate_jwt(room=jitsi_room, moderator=True, exp=end_date)
+
+        return MeetingData(
+            meeting_id=generate_uuid4(),
+            room_name=jitsi_room,
+            room_url=f"https://jitsi.domain/{jitsi_room}?jwt={user_jwt}",
+            host_room_url=f"https://jitsi.domain/{jitsi_room}?jwt={host_jwt}",
+            platform=self.PLATFORM_NAME,
+            extra_data={"user_jwt": user_jwt, "host_jwt": host_jwt}
+        )
+```
+
+### 2. **Settings Integration**
+
+```python
+# settings.py
+JITSI_DOMAIN: str = "meet.jit.si"
+JITSI_JWT_SECRET: str | None = None
+JITSI_WEBHOOK_SECRET: str | None = None
+JITSI_API_URL: str | None = None  # If using Jitsi API
+```
+
+### 3. **Factory Registration**
+
+```python
+# registry.py
+def _register_builtin_platforms():
+    from .jitsi import JitsiClient
+    register_platform("jitsi", JitsiClient)
+
+# factory.py
+def get_platform_config(platform: str) -> VideoPlatformConfig:
+    elif platform == "jitsi":
+        return VideoPlatformConfig(
+            api_key="",  # Jitsi may not need API key
+            webhook_secret=settings.JITSI_WEBHOOK_SECRET or "",
+            api_url=settings.JITSI_API_URL,
+        )
+```
+
+### 4. **Webhook Integration**
+
+```python
+# views/jitsi.py
+@router.post("/jitsi/events")
+async def jitsi_events_webhook(event_data: dict):
+    # Handle Prosody event-sync webhook format
+    event_type = event_data.get("event")
+    room_name = event_data.get("room", "").split("@")[0]
+
+    if event_type == "muc-occupant-joined":
+        # Same participant handling logic as other platforms
+```
+
+## Key Benefits of This Architecture
+
+### 1. **Isolation and Organization**
+- Platform-specific code contained in separate modules
+- No platform logic leaking into core application
+- Easy to add/remove platforms without affecting others
+
+### 2. **Consistent Interface**
+- All platforms implement the same abstract methods
+- Standardized `MeetingData` structure
+- Uniform error handling and logging
+
+### 3. **Gradual Migration Support**
+- Feature flags for controlled rollouts
+- Room-specific platform selection
+- Fallback mechanisms for platform failures
+
+### 4. **Configuration Management**
+- Centralized settings per platform
+- Consistent naming patterns
+- Environment-based configuration
+
+### 5. **Testing and Quality**
+- Mock platform for testing
+- Comprehensive test coverage
+- Platform-specific test utilities
+
+## Migration Strategy Applied
+
+The daily.co implementation demonstrates a careful migration approach:
+
+### 1. **Backward Compatibility**
+- Default platform remains "whereby"
+- Existing rooms continue using Whereby unless explicitly migrated
+- Same API endpoints and response formats
+
+### 2. **Feature Flag Control**
+```python
+# Gradual rollout control
+DAILY_MIGRATION_ENABLED: bool = True
+DAILY_MIGRATION_ROOM_IDS: list[str] = []  # Specific rooms to migrate
+DEFAULT_VIDEO_PLATFORM: str = "daily"     # New rooms default
+```
+
+### 3. **Data Integrity**
+- Platform field tracks which service each room/meeting uses
+- No data loss during migration
+- Platform-specific data preserved in `extra_data`
+
+### 4. **Monitoring and Rollback**
+- Comprehensive logging of platform selection
+- Easy rollback by changing feature flags
+- Platform-specific error tracking
+
+## Recommendations for Jitsi Integration
+
+Based on this analysis and the user's requirements:
+
+### 1. **Follow the Pattern**
+- Create `video_platforms/jitsi/` directory with:
+  - `client.py` - Main JitsiClient implementation
+  - `tasks.py` - Jitsi-specific worker tasks
+  - `__init__.py` - Module exports
+
+### 2. **Settings Organization**
+- Use `JITSI_*` prefix for all Jitsi settings
+- Follow the same configuration pattern as Daily.co
+- Support both environment variables and config files
+
+### 3. **Generic Database Fields**
+- Avoid platform-specific columns in database
+- Use `provider_data` JSON field if platform-specific data needed
+- Keep `platform` field as simple string identifier
+
+### 4. **Worker Task Migration**
+According to user requirements, migrate platform-specific tasks:
+```
+video_platforms/
+├── whereby/
+│   ├── client.py  (moved from whereby.py)
+│   └── tasks.py   (moved from worker/whereby_tasks.py)
+├── daily/
+│   ├── client.py  (moved from daily.py)
+│   └── tasks.py   (moved from worker/daily_tasks.py)
+└── jitsi/
+    ├── client.py  (new JitsiClient)
+    └── tasks.py   (new Jitsi recording tasks)
+```
+
+### 5. **Webhook Architecture**
+- Create `views/jitsi.py` for Jitsi-specific webhooks
+- Follow the same signature verification pattern
+- Reuse existing participant tracking logic
+
+## Implementation Checklist for Jitsi
+
+- [ ] Create `video_platforms/jitsi/` directory structure
+- [ ] Implement `JitsiClient` following the abstract interface
+- [ ] Add Jitsi settings to configuration
+- [ ] Register Jitsi platform in factory/registry
+- [ ] Create Jitsi webhook endpoint
+- [ ] Implement JWT token generation for room access
+- [ ] Add Jitsi recording processing tasks
+- [ ] Create comprehensive test suite
+- [ ] Update database migrations for platform field
+- [ ] Document Jitsi-specific configuration
+
+## Conclusion
+
+The video platforms refactoring in PR #529 provides an excellent foundation for adding Jitsi support. The architecture is well-designed with clear separation of concerns, consistent interfaces, and excellent extensibility. The daily.co implementation demonstrates how to add a new platform while maintaining backward compatibility and providing gradual migration capabilities.
+
+The pattern should be directly applicable to Jitsi integration, with the main differences being:
+- JWT-based authentication instead of API keys
+- Different webhook event formats
+- Jibri recording pipeline integration
+- Self-hosted deployment considerations
+
+This architecture successfully achieves the user's goals of:
+1. Settings-based configuration
+2. Generic database fields (no provider-specific columns)
+3. Platform isolation in separate directories
+4. Worker task organization within platform folders

server/contrib/jitsi/README.md

@@ -0,0 +1,212 @@
+# Event Logger for Docker-Jitsi-Meet
+
+A Prosody module that logs Jitsi meeting events to JSONL files alongside recordings, enabling complete participant tracking and speaker statistics.
+
+## Prerequisites
+
+- Running docker-jitsi-meet installation
+- Jibri configured for recording
+
+## Installation
+
+### Step 1: Copy the Module
+
+Copy the Prosody module to your custom plugins directory:
+
+```bash
+# Create the directory if it doesn't exist
+mkdir -p ~/.jitsi-meet-cfg/prosody/prosody-plugins-custom
+
+# Copy the module
+cp mod_event_logger.lua ~/.jitsi-meet-cfg/prosody/prosody-plugins-custom/
+```
+
+### Step 2: Update Your .env File
+
+Add or modify these variables in your `.env` file:
+
+```bash
+# If XMPP_MUC_MODULES already exists, append event_logger
+# Example: XMPP_MUC_MODULES=existing_module,event_logger
+XMPP_MUC_MODULES=event_logger
+
+# Optional: Configure the module (these are defaults)
+JIBRI_RECORDINGS_PATH=/config/recordings
+JIBRI_LOG_SPEAKER_STATS=true
+JIBRI_SPEAKER_STATS_INTERVAL=10
+```
+
+**Important**: If you already have `XMPP_MUC_MODULES` defined, add `event_logger` to the comma-separated list:
+```bash
+# Existing modules + our module
+XMPP_MUC_MODULES=mod_info,mod_alert,event_logger
+```
+
+### Step 3: Modify docker-compose.yml
+
+Add a shared recordings volume so Prosody can write events alongside Jibri recordings:
+
+```yaml
+services:
+  prosody:
+    # ... existing configuration ...
+    volumes:
+      - ${CONFIG}/prosody/config:/config:Z
+      - ${CONFIG}/prosody/prosody-plugins-custom:/prosody-plugins-custom:Z
+      - ${CONFIG}/recordings:/config/recordings:Z  # Add this line
+    environment:
+      # Add if not using .env file
+      - XMPP_MUC_MODULES=${XMPP_MUC_MODULES:-event_logger}
+      - JIBRI_RECORDINGS_PATH=/config/recordings
+
+  jibri:
+    # ... existing configuration ...
+    volumes:
+      - ${CONFIG}/jibri:/config:Z
+      - ${CONFIG}/recordings:/config/recordings:Z  # Add this line
+    environment:
+      # For Reflector webhook integration (optional)
+      - REFLECTOR_WEBHOOK_URL=${REFLECTOR_WEBHOOK_URL:-}
+      - JIBRI_FINALIZE_RECORDING_SCRIPT_PATH=/config/finalize.sh
+```
+
+### Step 4: Add Finalize Script (Optional - For Reflector Integration)
+
+If you want to notify Reflector when recordings complete:
+
+```bash
+# Copy the finalize script
+cp finalize.sh ~/.jitsi-meet-cfg/jibri/finalize.sh
+chmod +x ~/.jitsi-meet-cfg/jibri/finalize.sh
+
+# Add to .env
+REFLECTOR_WEBHOOK_URL=http://your-reflector-api:8000
+```
+
+### Step 5: Restart Services
+
+```bash
+docker-compose down
+docker-compose up -d
+```
+
+## What Gets Created
+
+After a recording, you'll find in `~/.jitsi-meet-cfg/recordings/{session-id}/`:
+- `recording.mp4` - The video recording (created by Jibri)
+- `metadata.json` - Basic metadata (created by Jibri)
+- `events.jsonl` - Complete participant timeline (created by this module)
+
+## Event Format
+
+Each line in `events.jsonl` is a JSON object:
+
+```json
+{"type":"room_created","timestamp":1234567890,"room_name":"TestRoom","room_jid":"testroom@conference.meet.jitsi","meeting_url":"https://meet.jitsi/TestRoom"}
+{"type":"recording_started","timestamp":1234567891,"room_name":"TestRoom","session_id":"20240115120000_TestRoom","jibri_jid":"jibri@recorder.meet.jitsi"}
+{"type":"participant_joined","timestamp":1234567892,"room_name":"TestRoom","participant":{"jid":"user1@meet.jitsi/web","nick":"John Doe","id":"user1@meet.jitsi","is_moderator":false}}
+{"type":"speaker_active","timestamp":1234567895,"room_name":"TestRoom","speaker_jid":"user1@meet.jitsi","speaker_nick":"John Doe","duration":10}
+{"type":"participant_left","timestamp":1234567920,"room_name":"TestRoom","participant":{"jid":"user1@meet.jitsi/web","nick":"John Doe","duration_seconds":28}}
+{"type":"recording_stopped","timestamp":1234567950,"room_name":"TestRoom","session_id":"20240115120000_TestRoom","meeting_url":"https://meet.jitsi/TestRoom"}
+```
+
+## Configuration Options
+
+All configuration can be done via environment variables:
+
+| Environment Variable | Default | Description |
+|---------------------|---------|-------------|
+| `JIBRI_RECORDINGS_PATH` | `/config/recordings` | Path where recordings are stored |
+| `JIBRI_LOG_SPEAKER_STATS` | `true` | Enable speaker statistics logging |
+| `JIBRI_SPEAKER_STATS_INTERVAL` | `10` | Seconds between speaker stats updates |
+
+## Verifying Installation
+
+Check that the module is loaded:
+```bash
+docker-compose logs prosody | grep "Event Logger"
+# Should see: "Event Logger loaded - writing to /config/recordings"
+```
+
+Check for events after a recording:
+```bash
+ls -la ~/.jitsi-meet-cfg/recordings/*/events.jsonl
+cat ~/.jitsi-meet-cfg/recordings/*/events.jsonl | jq .
+```
+
+## Troubleshooting
+
+### No events.jsonl file created
+
+1. **Check module is enabled**:
+   ```bash
+   docker-compose exec prosody grep -r "event_logger" /config
+   ```
+
+2. **Verify volume permissions**:
+   ```bash
+   docker-compose exec prosody ls -la /config/recordings
+   ```
+
+3. **Check Prosody logs for errors**:
+   ```bash
+   docker-compose logs prosody | grep -i error
+   ```
+
+### Module not loading
+
+1. **Verify file exists in container**:
+   ```bash
+   docker-compose exec prosody ls -la /prosody-plugins-custom/
+   ```
+
+2. **Check XMPP_MUC_MODULES format** (must be comma-separated, no spaces):
+   - ✅ Correct: `XMPP_MUC_MODULES=mod1,mod2,event_logger`
+   - ❌ Wrong: `XMPP_MUC_MODULES=mod1, mod2, event_logger`
+
+## Common docker-compose.yml Patterns
+
+### Minimal Addition (if you trust defaults)
+```yaml
+services:
+  prosody:
+    volumes:
+      - ${CONFIG}/recordings:/config/recordings:Z  # Just add this
+```
+
+### Full Configuration
+```yaml
+services:
+  prosody:
+    volumes:
+      - ${CONFIG}/prosody/config:/config:Z
+      - ${CONFIG}/prosody/prosody-plugins-custom:/prosody-plugins-custom:Z
+      - ${CONFIG}/recordings:/config/recordings:Z
+    environment:
+      - XMPP_MUC_MODULES=event_logger
+      - JIBRI_RECORDINGS_PATH=/config/recordings
+      - JIBRI_LOG_SPEAKER_STATS=true
+      - JIBRI_SPEAKER_STATS_INTERVAL=10
+
+  jibri:
+    volumes:
+      - ${CONFIG}/jibri:/config:Z
+      - ${CONFIG}/recordings:/config/recordings:Z
+    environment:
+      - JIBRI_RECORDING_DIR=/config/recordings
+      - JIBRI_FINALIZE_RECORDING_SCRIPT_PATH=/config/finalize.sh
+```
+
+## Integration with Reflector
+
+The finalize.sh script will automatically notify Reflector when a recording completes if `REFLECTOR_WEBHOOK_URL` is set. Reflector will receive:
+
+```json
+{
+  "session_id": "20240115120000_TestRoom",
+  "path": "20240115120000_TestRoom",
+  "meeting_url": "https://meet.jitsi/TestRoom"
+}
+```
+
+Reflector then processes the recording along with the complete participant timeline from `events.jsonl`.

server/contrib/jitsi/finalize.sh

@@ -0,0 +1,49 @@
+#!/bin/bash
+# Jibri finalize script to notify Reflector when recording is complete
+# This script is called by Jibri with the recording directory as argument
+
+RECORDING_PATH="$1"
+SESSION_ID=$(basename "$RECORDING_PATH")
+METADATA_FILE="$RECORDING_PATH/metadata.json"
+
+# Extract meeting URL from Jibri's metadata
+MEETING_URL=""
+if [ -f "$METADATA_FILE" ]; then
+    MEETING_URL=$(jq -r '.meeting_url' "$METADATA_FILE" 2>/dev/null || echo "")
+fi
+
+echo "[$(date)] Recording finalized: $RECORDING_PATH"
+echo "[$(date)] Session ID: $SESSION_ID"
+echo "[$(date)] Meeting URL: $MEETING_URL"
+
+# Check if events.jsonl was created by our Prosody module
+if [ -f "$RECORDING_PATH/events.jsonl" ]; then
+    EVENT_COUNT=$(wc -l < "$RECORDING_PATH/events.jsonl")
+    echo "[$(date)] Found events.jsonl with $EVENT_COUNT events"
+else
+    echo "[$(date)] Warning: No events.jsonl found"
+fi
+
+# Notify Reflector if webhook URL is configured
+if [ -n "$REFLECTOR_WEBHOOK_URL" ]; then
+    echo "[$(date)] Notifying Reflector at: $REFLECTOR_WEBHOOK_URL"
+
+    RESPONSE=$(curl -s -w "\n%{http_code}" -X POST "$REFLECTOR_WEBHOOK_URL/api/v1/jibri/recording-ready" \
+        -H "Content-Type: application/json" \
+        -d "{\"session_id\":\"$SESSION_ID\",\"path\":\"$SESSION_ID\",\"meeting_url\":\"$MEETING_URL\"}")
+
+    HTTP_CODE=$(echo "$RESPONSE" | tail -n1)
+    BODY=$(echo "$RESPONSE" | sed '$d')
+
+    if [ "$HTTP_CODE" = "200" ]; then
+        echo "[$(date)] Reflector notified successfully"
+        echo "[$(date)] Response: $BODY"
+    else
+        echo "[$(date)] Failed to notify Reflector. HTTP code: $HTTP_CODE"
+        echo "[$(date)] Response: $BODY"
+    fi
+else
+    echo "[$(date)] No REFLECTOR_WEBHOOK_URL configured, skipping notification"
+fi
+
+echo "[$(date)] Finalize script completed"

server/contrib/jitsi/mod_event_logger.lua

@@ -0,0 +1,372 @@
+local json = require "util.json"
+local st = require "util.stanza"
+local jid_bare = require "util.jid".bare
+
+local recordings_path = os.getenv("JIBRI_RECORDINGS_PATH") or
+                       module:get_option_string("jibri_recordings_path", "/recordings")
+
+-- room_jid -> { session_id, participants = {jid -> info} }
+local active_recordings = {}
+-- room_jid -> { participants = {jid -> info}, created_at }
+local room_states = {}
+
+local function get_timestamp()
+    return os.time()
+end
+
+local function write_event(session_id, event)
+    if not session_id then
+        module:log("warn", "No session_id for event: %s", event.type)
+        return
+    end
+
+    local session_dir = string.format("%s/%s", recordings_path, session_id)
+    local event_file = string.format("%s/events.jsonl", session_dir)
+
+    module:log("info", "Writing event %s to %s", event.type, event_file)
+
+    -- Create directory
+    local mkdir_cmd = string.format("mkdir -p '%s' 2>&1", session_dir)
+    local mkdir_result = os.execute(mkdir_cmd)
+    module:log("debug", "mkdir result: %s", tostring(mkdir_result))
+
+    local file, err = io.open(event_file, "a")
+    if file then
+        local json_str = json.encode(event)
+        file:write(json_str .. "\n")
+        file:close()
+        module:log("info", "Successfully wrote event %s", event.type)
+    else
+        module:log("error", "Failed to write event to %s: %s", event_file, err)
+    end
+end
+
+local function extract_participant_info(occupant)
+    local info = {
+        jid = occupant.jid,
+        bare_jid = occupant.bare_jid,
+        nick = occupant.nick,
+        display_name = nil,
+        role = occupant.role
+    }
+
+    local presence = occupant:get_presence()
+    if presence then
+        local nick_element = presence:get_child("nick", "http://jabber.org/protocol/nick")
+        if nick_element then
+            info.display_name = nick_element:get_text()
+        end
+
+        local identity = presence:get_child("identity")
+        if identity then
+            local user = identity:get_child("user")
+            if user then
+                local name = user:get_child("name")
+                if name then
+                    info.display_name = name:get_text()
+                end
+
+                local id_element = user:get_child("id")
+                if id_element then
+                    info.id = id_element:get_text()
+                end
+            end
+        end
+
+        if not info.display_name and occupant.nick then
+            local _, _, resource = occupant.nick:match("([^@]+)@([^/]+)/(.+)")
+            if resource then
+                info.display_name = resource
+            end
+        end
+    end
+
+    return info
+end
+
+local function get_room_participant_count(room)
+    local count = 0
+    for _ in room:each_occupant() do
+        count = count + 1
+    end
+    return count
+end
+
+local function snapshot_room_participants(room)
+    local participants = {}
+    local total = 0
+    local skipped = 0
+
+    module:log("info", "Snapshotting room participants")
+
+    for _, occupant in room:each_occupant() do
+        total = total + 1
+        -- Skip recorders (Jibri)
+        if occupant.bare_jid and (occupant.bare_jid:match("^recorder@") or
+                                  occupant.bare_jid:match("^jibri@")) then
+            skipped = skipped + 1
+        else
+            local info = extract_participant_info(occupant)
+            participants[occupant.jid] = info
+            module:log("debug", "Added participant: %s", info.display_name or info.bare_jid)
+        end
+    end
+
+    module:log("info", "Snapshot: %d total, %d participants", total, total - skipped)
+    return participants
+end
+
+-- Import utility functions if available
+local util = module:require "util";
+local get_room_from_jid = util.get_room_from_jid;
+local room_jid_match_rewrite = util.room_jid_match_rewrite;
+
+-- Main IQ handler for Jibri stanzas
+module:hook("pre-iq/full", function(event)
+    local stanza = event.stanza
+    if stanza.name ~= "iq" then
+        return
+    end
+
+    local jibri = stanza:get_child('jibri', 'http://jitsi.org/protocol/jibri')
+    if not jibri then
+        return
+    end
+
+    module:log("info", "=== Jibri IQ intercepted ===")
+
+    local action = jibri.attr.action
+    local session_id = jibri.attr.session_id
+    local room_jid = jibri.attr.room
+    local recording_mode = jibri.attr.recording_mode
+    local app_data = jibri.attr.app_data
+
+    module:log("info", "Jibri %s - session: %s, room: %s, mode: %s",
+               action or "?", session_id or "?", room_jid or "?", recording_mode or "?")
+
+    if not room_jid or not session_id then
+        module:log("warn", "Missing room_jid or session_id")
+        return
+    end
+
+    -- Get the room using util function
+    local room = get_room_from_jid(room_jid_match_rewrite(jid_bare(stanza.attr.to)))
+    if not room then
+        -- Try with the room_jid directly
+        room = get_room_from_jid(room_jid)
+    end
+
+    if not room then
+        module:log("error", "Room not found for jid: %s", room_jid)
+        return
+    end
+
+    module:log("info", "Room found: %s", room:get_name() or room_jid)
+
+    if action == "start" then
+        module:log("info", "Recording START for session %s", session_id)
+
+        -- Count and snapshot participants
+        local participant_count = 0
+        for _ in room:each_occupant() do
+            participant_count = participant_count + 1
+        end
+
+        local participants = snapshot_room_participants(room)
+        local participant_list = {}
+        for jid, info in pairs(participants) do
+            table.insert(participant_list, info)
+        end
+
+        active_recordings[room_jid] = {
+            session_id = session_id,
+            participants = participants,
+            started_at = get_timestamp()
+        }
+
+        write_event(session_id, {
+            type = "recording_started",
+            timestamp = get_timestamp(),
+            room_jid = room_jid,
+            room_name = room:get_name(),
+            session_id = session_id,
+            recording_mode = recording_mode,
+            app_data = app_data,
+            participant_count = participant_count,
+            participants_at_start = participant_list
+        })
+
+    elseif action == "stop" then
+        module:log("info", "Recording STOP for session %s", session_id)
+
+        local recording = active_recordings[room_jid]
+        if recording and recording.session_id == session_id then
+            write_event(session_id, {
+                type = "recording_stopped",
+                timestamp = get_timestamp(),
+                room_jid = room_jid,
+                room_name = room:get_name(),
+                session_id = session_id,
+                duration = get_timestamp() - recording.started_at,
+                participant_count = get_room_participant_count(room)
+            })
+
+            active_recordings[room_jid] = nil
+        else
+            module:log("warn", "No active recording found for room %s", room_jid)
+        end
+    end
+end);
+
+-- Room and participant event hooks
+local function setup_room_hooks(host_module)
+    module:log("info", "Setting up room hooks on %s", host_module.host or "unknown")
+
+    -- Room created
+    host_module:hook("muc-room-created", function(event)
+        local room = event.room
+        local room_jid = room.jid
+
+        room_states[room_jid] = {
+            participants = {},
+            created_at = get_timestamp()
+        }
+
+        module:log("info", "Room created: %s", room_jid)
+    end)
+
+    -- Room destroyed
+    host_module:hook("muc-room-destroyed", function(event)
+        local room = event.room
+        local room_jid = room.jid
+
+        room_states[room_jid] = nil
+        active_recordings[room_jid] = nil
+
+        module:log("info", "Room destroyed: %s", room_jid)
+    end)
+
+    -- Occupant joined
+    host_module:hook("muc-occupant-joined", function(event)
+        local room = event.room
+        local occupant = event.occupant
+        local room_jid = room.jid
+
+        -- Skip recorders
+        if occupant.bare_jid and (occupant.bare_jid:match("^recorder@") or
+                                   occupant.bare_jid:match("^jibri@")) then
+            return
+        end
+
+        local participant_info = extract_participant_info(occupant)
+
+        -- Update room state
+        if room_states[room_jid] then
+            room_states[room_jid].participants[occupant.jid] = participant_info
+        end
+
+        -- Log to active recording if exists
+        local recording = active_recordings[room_jid]
+        if recording then
+            recording.participants[occupant.jid] = participant_info
+
+            write_event(recording.session_id, {
+                type = "participant_joined",
+                timestamp = get_timestamp(),
+                room_jid = room_jid,
+                room_name = room:get_name(),
+                participant = participant_info,
+                participant_count = get_room_participant_count(room)
+            })
+        end
+
+        module:log("info", "Participant joined %s: %s (%d total)",
+                   room:get_name() or room_jid,
+                   participant_info.display_name or participant_info.bare_jid,
+                   get_room_participant_count(room))
+    end)
+
+    -- Occupant left
+    host_module:hook("muc-occupant-left", function(event)
+        local room = event.room
+        local occupant = event.occupant
+        local room_jid = room.jid
+
+        -- Skip recorders
+        if occupant.bare_jid and (occupant.bare_jid:match("^recorder@") or
+                                   occupant.bare_jid:match("^jibri@")) then
+            return
+        end
+
+        local participant_info = extract_participant_info(occupant)
+
+        -- Update room state
+        if room_states[room_jid] then
+            room_states[room_jid].participants[occupant.jid] = nil
+        end
+
+        -- Log to active recording if exists
+        local recording = active_recordings[room_jid]
+        if recording then
+            if recording.participants[occupant.jid] then
+                recording.participants[occupant.jid] = nil
+            end
+
+            write_event(recording.session_id, {
+                type = "participant_left",
+                timestamp = get_timestamp(),
+                room_jid = room_jid,
+                room_name = room:get_name(),
+                participant = participant_info,
+                participant_count = get_room_participant_count(room)
+            })
+        end
+
+        module:log("info", "Participant left %s: %s (%d remaining)",
+                   room:get_name() or room_jid,
+                   participant_info.display_name or participant_info.bare_jid,
+                   get_room_participant_count(room))
+    end)
+end
+
+-- Module initialization
+local current_host = module:get_host()
+local host_type = module:get_host_type()
+
+module:log("info", "Event Logger loading on %s (type: %s)", current_host, host_type or "unknown")
+module:log("info", "Recording path: %s", recordings_path)
+
+-- Setup room hooks based on host type
+if host_type == "component" and current_host:match("^[^.]+%.") then
+    setup_room_hooks(module)
+else
+    -- Try to find and hook to MUC component
+    local process_host_module = util.process_host_module
+    local muc_component_host = module:get_option_string("muc_component") or
+                              module:get_option_string("main_muc")
+
+    if not muc_component_host then
+        local possible_hosts = {
+            "muc." .. current_host,
+            "conference." .. current_host,
+            "rooms." .. current_host
+        }
+
+        for _, host in ipairs(possible_hosts) do
+            if prosody.hosts[host] then
+                muc_component_host = host
+                module:log("info", "Auto-detected MUC component: %s", muc_component_host)
+                break
+            end
+        end
+    end
+
+    if muc_component_host then
+        process_host_module(muc_component_host, function(host_module, host)
+            module:log("info", "Hooking to MUC events on %s", host)
+            setup_room_hooks(host_module)
+        end)
+    else
+        module:log("error", "Could not find MUC component")
+    end
+end

server/docs/data_retention.md

@@ -0,0 +1,95 @@
+# Data Retention and Cleanup
+
+## Overview
+
+For public instances of Reflector, a data retention policy is automatically enforced to delete anonymous user data after a configurable period (default: 7 days). This ensures compliance with privacy expectations and prevents unbounded storage growth.
+
+## Configuration
+
+### Environment Variables
+
+- `PUBLIC_MODE` (bool): Must be set to `true` to enable automatic cleanup
+- `PUBLIC_DATA_RETENTION_DAYS` (int): Number of days to retain anonymous data (default: 7)
+
+### What Gets Deleted
+
+When data reaches the retention period, the following items are automatically removed:
+
+1. **Transcripts** from anonymous users (where `user_id` is NULL):
+   - Database records
+   - Local files (audio.wav, audio.mp3, audio.json waveform)
+   - Storage files (cloud storage if configured)
+
+## Automatic Cleanup
+
+### Celery Beat Schedule
+
+When `PUBLIC_MODE=true`, a Celery beat task runs daily at 3 AM to clean up old data:
+
+```python
+# Automatically scheduled when PUBLIC_MODE=true
+"cleanup_old_public_data": {
+    "task": "reflector.worker.cleanup.cleanup_old_public_data",
+    "schedule": crontab(hour=3, minute=0),  # Daily at 3 AM
+}
+```
+
+### Running the Worker
+
+Ensure both Celery worker and beat scheduler are running:
+
+```bash
+# Start Celery worker
+uv run celery -A reflector.worker.app worker --loglevel=info
+
+# Start Celery beat scheduler (in another terminal)
+uv run celery -A reflector.worker.app beat
+```
+
+## Manual Cleanup
+
+For testing or manual intervention, use the cleanup tool:
+
+```bash
+# Delete data older than 7 days (default)
+uv run python -m reflector.tools.cleanup_old_data
+
+# Delete data older than 30 days
+uv run python -m reflector.tools.cleanup_old_data --days 30
+```
+
+Note: The manual tool uses the same implementation as the Celery worker task to ensure consistency.
+
+## Important Notes
+
+1. **User Data Deletion**: Only anonymous data (where `user_id` is NULL) is deleted. Authenticated user data is preserved.
+
+2. **Storage Cleanup**: The system properly cleans up both local files and cloud storage when configured.
+
+3. **Error Handling**: If individual deletions fail, the cleanup continues and logs errors. Failed deletions are reported in the task output.
+
+4. **Public Instance Only**: The automatic cleanup task only runs when `PUBLIC_MODE=true` to prevent accidental data loss in private deployments.
+
+## Testing
+
+Run the cleanup tests:
+
+```bash
+uv run pytest tests/test_cleanup.py -v
+```
+
+## Monitoring
+
+Check Celery logs for cleanup task execution:
+
+```bash
+# Look for cleanup task logs
+grep "cleanup_old_public_data" celery.log
+grep "Starting cleanup of old public data" celery.log
+```
+
+Task statistics are logged after each run:
+- Number of transcripts deleted
+- Number of meetings deleted
+- Number of orphaned recordings deleted
+- Any errors encountered

server/docs/gpu/api-transcription.md

@@ -0,0 +1,194 @@
+## Reflector GPU Transcription API (Specification)
+
+This document defines the Reflector GPU transcription API that all implementations must adhere to. Current implementations include NVIDIA Parakeet (NeMo) and Whisper (faster-whisper), both deployed on Modal.com. The API surface and response shapes are OpenAI/Whisper-compatible, so clients can switch implementations by changing only the base URL.
+
+### Base URL and Authentication
+
+- Example base URLs (Modal web endpoints):
+
+  - Parakeet: `https://<account>--reflector-transcriber-parakeet-web.modal.run`
+  - Whisper: `https://<account>--reflector-transcriber-web.modal.run`
+
+- All endpoints are served under `/v1` and require a Bearer token:
+
+```
+Authorization: Bearer <REFLECTOR_GPU_APIKEY>
+```
+
+Note: To switch implementations, deploy the desired variant and point `TRANSCRIPT_URL` to its base URL. The API is identical.
+
+### Supported file types
+
+`mp3, mp4, mpeg, mpga, m4a, wav, webm`
+
+### Models and languages
+
+- Parakeet (NVIDIA NeMo): default `nvidia/parakeet-tdt-0.6b-v2`
+  - Language support: only `en`. Other languages return HTTP 400.
+- Whisper (faster-whisper): default `large-v2` (or deployment-specific)
+  - Language support: multilingual (per Whisper model capabilities).
+
+Note: The `model` parameter is accepted by all implementations for interface parity. Some backends may treat it as informational.
+
+### Endpoints
+
+#### POST /v1/audio/transcriptions
+
+Transcribe one or more uploaded audio files.
+
+Request: multipart/form-data
+
+- `file` (File) — optional. Single file to transcribe.
+- `files` (File[]) — optional. One or more files to transcribe.
+- `model` (string) — optional. Defaults to the implementation-specific model (see above).
+- `language` (string) — optional, defaults to `en`.
+  - Parakeet: only `en` is accepted; other values return HTTP 400
+  - Whisper: model-dependent; typically multilingual
+- `batch` (boolean) — optional, defaults to `false`.
+
+Notes:
+
+- Provide either `file` or `files`, not both. If neither is provided, HTTP 400.
+- `batch` requires `files`; using `batch=true` without `files` returns HTTP 400.
+- Response shape for multiple files is the same regardless of `batch`.
+- Files sent to this endpoint are processed in a single pass (no VAD/chunking). This is intended for short clips (roughly ≤ 30s; depends on GPU memory/model). For longer audio, prefer `/v1/audio/transcriptions-from-url` which supports VAD-based chunking.
+
+Responses
+
+Single file response:
+
+```json
+{
+  "text": "transcribed text",
+  "words": [
+    { "word": "hello", "start": 0.0, "end": 0.5 },
+    { "word": "world", "start": 0.5, "end": 1.0 }
+  ],
+  "filename": "audio.mp3"
+}
+```
+
+Multiple files response:
+
+```json
+{
+  "results": [
+    {"filename": "a1.mp3", "text": "...", "words": [...]},
+    {"filename": "a2.mp3", "text": "...", "words": [...]}]
+}
+```
+
+Notes:
+
+- Word objects always include keys: `word`, `start`, `end`.
+- Some implementations may include a trailing space in `word` to match Whisper tokenization behavior; clients should trim if needed.
+
+Example curl (single file):
+
+```bash
+curl -X POST \
+  -H "Authorization: Bearer $REFLECTOR_GPU_APIKEY" \
+  -F "file=@/path/to/audio.mp3" \
+  -F "language=en" \
+  "$BASE_URL/v1/audio/transcriptions"
+```
+
+Example curl (multiple files, batch):
+
+```bash
+curl -X POST \
+  -H "Authorization: Bearer $REFLECTOR_GPU_APIKEY" \
+  -F "files=@/path/a1.mp3" -F "files=@/path/a2.mp3" \
+  -F "batch=true" -F "language=en" \
+  "$BASE_URL/v1/audio/transcriptions"
+```
+
+#### POST /v1/audio/transcriptions-from-url
+
+Transcribe a single remote audio file by URL.
+
+Request: application/json
+
+Body parameters:
+
+- `audio_file_url` (string) — required. URL of the audio file to transcribe.
+- `model` (string) — optional. Defaults to the implementation-specific model (see above).
+- `language` (string) — optional, defaults to `en`. Parakeet only accepts `en`.
+- `timestamp_offset` (number) — optional, defaults to `0.0`. Added to each word's `start`/`end` in the response.
+
+```json
+{
+  "audio_file_url": "https://example.com/audio.mp3",
+  "model": "nvidia/parakeet-tdt-0.6b-v2",
+  "language": "en",
+  "timestamp_offset": 0.0
+}
+```
+
+Response:
+
+```json
+{
+  "text": "transcribed text",
+  "words": [
+    { "word": "hello", "start": 10.0, "end": 10.5 },
+    { "word": "world", "start": 10.5, "end": 11.0 }
+  ]
+}
+```
+
+Notes:
+
+- `timestamp_offset` is added to each word’s `start`/`end` in the response.
+- Implementations may perform VAD-based chunking and batching for long-form audio; word timings are adjusted accordingly.
+
+Example curl:
+
+```bash
+curl -X POST \
+  -H "Authorization: Bearer $REFLECTOR_GPU_APIKEY" \
+  -H "Content-Type: application/json" \
+  -d '{
+        "audio_file_url": "https://example.com/audio.mp3",
+        "language": "en",
+        "timestamp_offset": 0
+      }' \
+  "$BASE_URL/v1/audio/transcriptions-from-url"
+```
+
+### Error handling
+
+- 400 Bad Request
+  - Parakeet: `language` other than `en`
+  - Missing required parameters (`file`/`files` for upload; `audio_file_url` for URL endpoint)
+  - Unsupported file extension
+- 401 Unauthorized
+  - Missing or invalid Bearer token
+- 404 Not Found
+  - `audio_file_url` does not exist
+
+### Implementation details
+
+- GPUs: A10G for small-file/live, L40S for large-file URL transcription (subject to deployment)
+- VAD chunking and segment batching; word timings adjusted and overlapping ends constrained
+- Pads very short segments (< 0.5s) to avoid model crashes on some backends
+
+### Server configuration (Reflector API)
+
+Set the Reflector server to use the Modal backend and point `TRANSCRIPT_URL` to your chosen deployment:
+
+```
+TRANSCRIPT_BACKEND=modal
+TRANSCRIPT_URL=https://<account>--reflector-transcriber-parakeet-web.modal.run
+TRANSCRIPT_MODAL_API_KEY=<REFLECTOR_GPU_APIKEY>
+```
+
+### Conformance tests
+
+Use the pytest-based conformance tests to validate any new implementation (including self-hosted) against this spec:
+
+```
+TRANSCRIPT_URL=https://<your-deployment-base> \
+TRANSCRIPT_MODAL_API_KEY=your-api-key \
+uv run -m pytest -m gpu_modal --no-cov server/tests/test_gpu_modal_transcript.py
+```

server/docs/platform-jitsi.md

@@ -0,0 +1,493 @@
+# Jitsi Integration Configuration Guide
+
+This guide provides step-by-step instructions for configuring Reflector to work with a self-hosted Jitsi Meet installation for video meetings and recording.
+
+## Prerequisites
+
+Before configuring Jitsi integration, ensure you have:
+
+- **Self-hosted Jitsi Meet installation** (version 2.0.8922 or later recommended)
+- **Jibri recording service** configured and running
+- **Prosody XMPP server** with mod_event_sync module installed
+- **Docker or system deployment** of Reflector with access to environment variables
+- **SSL certificates** for secure communication between services
+
+## Environment Configuration
+
+Add the following environment variables to your Reflector deployment:
+
+### Required Settings
+
+```bash
+# Jitsi Meet domain (without https://)
+JITSI_DOMAIN=meet.example.com
+
+# JWT secret for room authentication (generate with: openssl rand -hex 32)
+JITSI_JWT_SECRET=your-64-character-hex-secret-here
+
+# Webhook secret for secure event handling (generate with: openssl rand -hex 16)
+JITSI_WEBHOOK_SECRET=your-32-character-hex-secret-here
+
+# Application identifier (should match Jitsi configuration)
+JITSI_APP_ID=reflector
+
+# JWT issuer and audience (should match Jitsi configuration)
+JITSI_JWT_ISSUER=reflector
+JITSI_JWT_AUDIENCE=jitsi
+```
+
+### Example .env Configuration
+
+```bash
+# Add to your server/.env file
+JITSI_DOMAIN=meet.mycompany.com
+JITSI_JWT_SECRET=$(openssl rand -hex 32)
+JITSI_WEBHOOK_SECRET=$(openssl rand -hex 16)
+JITSI_APP_ID=reflector
+JITSI_JWT_ISSUER=reflector
+JITSI_JWT_AUDIENCE=jitsi
+```
+
+## Jitsi Meet Server Configuration
+
+### 1. JWT Authentication Setup
+
+Edit `/etc/prosody/conf.d/[YOUR_DOMAIN].cfg.lua`:
+
+```lua
+VirtualHost "meet.example.com"
+    authentication = "token"
+    app_id = "reflector"
+    app_secret = "your-jwt-secret-here"
+
+    -- Allow anonymous access for non-authenticated users
+    c2s_require_encryption = false
+    admins = { "focusUser@auth.meet.example.com" }
+
+    modules_enabled = {
+        "bosh";
+        "pubsub";
+        "ping";
+        "roster";
+        "saslauth";
+        "tls";
+        "dialback";
+        "disco";
+        "carbons";
+        "pep";
+        "private";
+        "blocklist";
+        "vcard";
+        "version";
+        "uptime";
+        "time";
+        "ping";
+        "register";
+        "admin_adhoc";
+        "token_verification";
+        "event_sync";  -- Required for webhook events
+    }
+```
+
+### 2. Room Access Control
+
+Edit `/etc/jitsi/meet/meet.example.com-config.js`:
+
+```javascript
+var config = {
+    hosts: {
+        domain: 'meet.example.com',
+        muc: 'conference.meet.example.com'
+    },
+
+    // Enable JWT authentication
+    enableUserRolesBasedOnToken: true,
+
+    // Recording configuration
+    fileRecordingsEnabled: true,
+    liveStreamingEnabled: false,
+
+    // Reflector-specific settings
+    prejoinPageEnabled: true,
+    requireDisplayName: true,
+};
+```
+
+### 3. Interface Configuration
+
+Edit `/usr/share/jitsi-meet/interface_config.js`:
+
+```javascript
+var interfaceConfig = {
+    // Customize for Reflector branding
+    APP_NAME: 'Reflector Meeting',
+    DEFAULT_WELCOME_PAGE_LOGO_URL: 'https://your-domain.com/logo.png',
+
+    // Hide unnecessary buttons
+    TOOLBAR_BUTTONS: [
+        'microphone', 'camera', 'closedcaptions', 'desktop',
+        'fullscreen', 'fodeviceselection', 'hangup',
+        'chat', 'recording', 'livestreaming', 'etherpad',
+        'sharedvideo', 'settings', 'raisehand', 'videoquality',
+        'filmstrip', 'invite', 'feedback', 'stats', 'shortcuts',
+        'tileview', 'videobackgroundblur', 'download', 'help',
+        'mute-everyone'
+    ]
+};
+```
+
+## Jibri Configuration
+
+### 1. Recording Service Setup
+
+Edit `/etc/jitsi/jibri/jibri.conf`:
+
+```hocon
+jibri {
+    recording {
+        recordings-directory = "/var/recordings"
+        finalize-script = "/opt/jitsi/jibri/finalize.sh"
+    }
+
+    api {
+        xmpp {
+            environments = [{
+                name = "prod environment"
+                xmpp-server-hosts = ["meet.example.com"]
+                xmpp-domain = "meet.example.com"
+
+                control-muc {
+                    domain = "internal.auth.meet.example.com"
+                    room-name = "JibriBrewery"
+                    nickname = "jibri-nickname"
+                }
+
+                control-login {
+                    domain = "auth.meet.example.com"
+                    username = "jibri"
+                    password = "jibri-password"
+                }
+            }]
+        }
+    }
+}
+```
+
+### 2. Finalize Script Setup
+
+Create `/opt/jitsi/jibri/finalize.sh`:
+
+```bash
+#!/bin/bash
+# Jibri finalize script for Reflector integration
+
+RECORDING_FILE="$1"
+ROOM_NAME="$2"
+REFLECTOR_API_URL="${REFLECTOR_API_URL:-http://localhost:1250}"
+WEBHOOK_SECRET="${JITSI_WEBHOOK_SECRET}"
+
+# Generate webhook signature
+generate_signature() {
+    local payload="$1"
+    echo -n "$payload" | openssl dgst -sha256 -hmac "$WEBHOOK_SECRET" | cut -d' ' -f2
+}
+
+# Prepare webhook payload
+TIMESTAMP=$(date -u +%Y-%m-%dT%H:%M:%S.%3NZ)
+PAYLOAD=$(cat <<EOF
+{
+    "room_name": "$ROOM_NAME",
+    "recording_file": "$RECORDING_FILE",
+    "recording_status": "completed",
+    "timestamp": "$TIMESTAMP"
+}
+EOF
+)
+
+# Generate signature
+SIGNATURE=$(generate_signature "$PAYLOAD")
+
+# Send webhook to Reflector
+curl -X POST "$REFLECTOR_API_URL/v1/jibri/recording-complete" \
+    -H "Content-Type: application/json" \
+    -H "X-Jitsi-Signature: $SIGNATURE" \
+    -d "$PAYLOAD" \
+    --max-time 30
+
+echo "Recording finalization webhook sent for room: $ROOM_NAME"
+```
+
+Make the script executable:
+
+```bash
+chmod +x /opt/jitsi/jibri/finalize.sh
+```
+
+## Prosody Event Configuration
+
+### 1. Event-Sync Module Installation
+
+Install the mod_event_sync module:
+
+```bash
+# Download the module
+cd /usr/share/jitsi-meet/prosody-plugins/
+wget https://raw.githubusercontent.com/jitsi-contrib/prosody-plugins/main/mod_event_sync.lua
+
+# Or if using git
+git clone https://github.com/jitsi-contrib/prosody-plugins.git
+cp prosody-plugins/mod_event_sync.lua /usr/share/jitsi-meet/prosody-plugins/
+```
+
+### 2. Webhook Configuration
+
+Add to `/etc/prosody/conf.d/[YOUR_DOMAIN].cfg.lua`:
+
+```lua
+Component "conference.meet.example.com" "muc"
+    storage = "memory"
+    modules_enabled = {
+        "muc_meeting_id";
+        "muc_domain_mapper";
+        "polls";
+        "event_sync";  -- Enable event sync
+    }
+
+    -- Event sync webhook configuration
+    event_sync_url = "https://your-reflector-domain.com/v1/jitsi/events"
+    event_sync_secret = "your-webhook-secret-here"
+
+    -- Events to track
+    event_sync_events = {
+        "muc-occupant-joined",
+        "muc-occupant-left",
+        "jibri-recording-on",
+        "jibri-recording-off"
+    }
+```
+
+### 3. Restart Services
+
+After configuration changes, restart all services:
+
+```bash
+systemctl restart prosody
+systemctl restart jicofo
+systemctl restart jitsi-videobridge2
+systemctl restart jibri
+systemctl restart nginx
+```
+
+## Reflector Room Configuration
+
+### 1. Create Jitsi Room
+
+When creating rooms in Reflector, set the platform field:
+
+```bash
+curl -X POST "https://your-reflector-domain.com/v1/rooms" \
+    -H "Authorization: Bearer $AUTH_TOKEN" \
+    -H "Content-Type: application/json" \
+    -d '{
+        "name": "my-jitsi-room",
+        "platform": "jitsi",
+        "recording_type": "cloud",
+        "recording_trigger": "automatic-2nd-participant",
+        "is_locked": false,
+        "room_mode": "normal"
+    }'
+```
+
+### 2. Meeting Creation
+
+Meetings will automatically use Jitsi when the room platform is set to "jitsi":
+
+```bash
+curl -X POST "https://your-reflector-domain.com/v1/rooms/my-jitsi-room/meeting" \
+    -H "Authorization: Bearer $AUTH_TOKEN"
+```
+
+## Testing the Integration
+
+### 1. Health Check
+
+Verify Jitsi webhook configuration:
+
+```bash
+curl "https://your-reflector-domain.com/v1/jitsi/health"
+```
+
+Expected response:
+```json
+{
+    "status": "ok",
+    "service": "jitsi-webhooks",
+    "timestamp": "2025-01-15T10:30:00.000Z",
+    "webhook_secret_configured": true
+}
+```
+
+### 2. Room Creation Test
+
+1. Create a Jitsi room via Reflector API
+2. Start a meeting - should generate Jitsi Meet URL with JWT token
+3. Join with multiple participants - should trigger participant events
+4. Start recording - should trigger Jibri recording workflow
+
+### 3. Webhook Event Test
+
+Monitor Reflector logs for incoming webhook events:
+
+```bash
+# Check for participant events
+curl -X POST "https://your-reflector-domain.com/v1/jitsi/events" \
+    -H "Content-Type: application/json" \
+    -H "X-Jitsi-Signature: test-signature" \
+    -d '{
+        "event": "muc-occupant-joined",
+        "room": "test-room-name",
+        "timestamp": "2025-01-15T10:30:00.000Z",
+        "data": {}
+    }'
+```
+
+## Troubleshooting
+
+### Common Issues
+
+#### JWT Authentication Failures
+
+**Symptoms:** Users can't join rooms, "Authentication failed" errors
+
+**Solutions:**
+1. Verify JWT secret matches between Jitsi and Reflector
+2. Check JWT token expiration (default 8 hours)
+3. Ensure system clocks are synchronized
+4. Validate JWT issuer/audience configuration
+
+```bash
+# Debug JWT tokens
+echo "JWT_TOKEN_HERE" | cut -d'.' -f2 | base64 -d | jq
+```
+
+#### Webhook Events Not Received
+
+**Symptoms:** Participant counts not updating, recording events missing
+
+**Solutions:**
+1. Verify event_sync module is loaded in Prosody
+2. Check webhook URL accessibility from Jitsi server
+3. Validate webhook signature generation
+4. Review Prosody and Reflector logs
+
+```bash
+# Test webhook connectivity
+curl -v "https://your-reflector-domain.com/v1/jitsi/health"
+
+# Check Prosody logs
+tail -f /var/log/prosody/prosody.log
+
+# Check Reflector logs
+docker logs your-reflector-container
+```
+
+#### Recording Issues
+
+**Symptoms:** Recordings not starting, finalize script errors
+
+**Solutions:**
+1. Verify Jibri service status and configuration
+2. Check recording directory permissions
+3. Validate finalize script execution permissions
+4. Monitor Jibri logs for errors
+
+```bash
+# Check Jibri status
+systemctl status jibri
+
+# Test finalize script
+sudo -u jibri /opt/jitsi/jibri/finalize.sh "/test/recording.mp4" "test-room"
+
+# Check Jibri logs
+journalctl -u jibri -f
+```
+
+### Debug Commands
+
+```bash
+# Verify Jitsi configuration
+prosodyctl check config
+
+# Test JWT generation
+curl -X POST "https://your-reflector-domain.com/v1/rooms/test/meeting" \
+    -H "Authorization: Bearer $TOKEN" -v
+
+# Monitor webhook events
+tail -f /var/log/reflector/app.log | grep jitsi
+
+# Check room participant counts
+curl "https://your-reflector-domain.com/v1/rooms" \
+    -H "Authorization: Bearer $TOKEN" | jq '.data[].num_clients'
+```
+
+### Performance Optimization
+
+#### For High-Concurrent Usage
+
+1. **Jitsi Videobridge Tuning:**
+```bash
+# /etc/jitsi/videobridge/sip-communicator.properties
+org.jitsi.videobridge.STATISTICS_INTERVAL=5000
+org.jitsi.videobridge.load.INITIAL_STREAM_LIMIT=50
+```
+
+2. **Database Connection Pooling:**
+```python
+# In your Reflector settings
+DATABASE_POOL_SIZE=20
+DATABASE_MAX_OVERFLOW=30
+```
+
+3. **Redis Configuration:**
+```bash
+# For webhook event caching
+REDIS_URL=redis://localhost:6379/1
+WEBHOOK_EVENT_TTL=3600
+```
+
+## Security Considerations
+
+### Network Security
+- Use HTTPS/WSS for all communications
+- Implement proper firewall rules
+- Consider VPN for server-to-server communication
+
+### Authentication Security
+- Rotate JWT secrets regularly
+- Use strong webhook secrets (32+ characters)
+- Implement rate limiting on webhook endpoints
+
+### Recording Security
+- Encrypt recordings at rest
+- Implement access controls for recording files
+- Regular security audits of file permissions
+
+## Support
+
+For additional support:
+
+1. **Reflector Issues:** Check GitHub issues or create new ones
+2. **Jitsi Community:** [Community Forum](https://community.jitsi.org/)
+3. **Documentation:** [Jitsi Developer Guide](https://jitsi.github.io/handbook/)
+
+## Migration from Whereby
+
+If migrating from Whereby integration:
+
+1. Update existing rooms to use "jitsi" platform
+2. Verify webhook configurations are updated
+3. Test recording workflows thoroughly
+4. Monitor participant event accuracy
+5. Update any custom integrations using meeting APIs
+
+The platform abstraction layer ensures smooth migration with minimal API changes.

server/docs/webhook.md

@@ -0,0 +1,212 @@
+# Reflector Webhook Documentation
+
+## Overview
+
+Reflector supports webhook notifications to notify external systems when transcript processing is completed. Webhooks can be configured per room and are triggered automatically after a transcript is successfully processed.
+
+## Configuration
+
+Webhooks are configured at the room level with two fields:
+- `webhook_url`: The HTTPS endpoint to receive webhook notifications
+- `webhook_secret`: Optional secret key for HMAC signature verification (auto-generated if not provided)
+
+## Events
+
+### `transcript.completed`
+
+Triggered when a transcript has been fully processed, including transcription, diarization, summarization, and topic detection.
+
+### `test`
+
+A test event that can be triggered manually to verify webhook configuration.
+
+## Webhook Request Format
+
+### Headers
+
+All webhook requests include the following headers:
+
+| Header | Description | Example |
+|--------|-------------|---------|
+| `Content-Type` | Always `application/json` | `application/json` |
+| `User-Agent` | Identifies Reflector as the source | `Reflector-Webhook/1.0` |
+| `X-Webhook-Event` | The event type | `transcript.completed` or `test` |
+| `X-Webhook-Retry` | Current retry attempt number | `0`, `1`, `2`... |
+| `X-Webhook-Signature` | HMAC signature (if secret configured) | `t=1735306800,v1=abc123...` |
+
+### Signature Verification
+
+If a webhook secret is configured, Reflector includes an HMAC-SHA256 signature in the `X-Webhook-Signature` header to verify the webhook authenticity.
+
+The signature format is: `t={timestamp},v1={signature}`
+
+To verify the signature:
+1. Extract the timestamp and signature from the header
+2. Create the signed payload: `{timestamp}.{request_body}`
+3. Compute HMAC-SHA256 of the signed payload using your webhook secret
+4. Compare the computed signature with the received signature
+
+Example verification (Python):
+```python
+import hmac
+import hashlib
+
+def verify_webhook_signature(payload: bytes, signature_header: str, secret: str) -> bool:
+    # Parse header: "t=1735306800,v1=abc123..."
+    parts = dict(part.split("=") for part in signature_header.split(","))
+    timestamp = parts["t"]
+    received_signature = parts["v1"]
+
+    # Create signed payload
+    signed_payload = f"{timestamp}.{payload.decode('utf-8')}"
+
+    # Compute expected signature
+    expected_signature = hmac.new(
+        secret.encode("utf-8"),
+        signed_payload.encode("utf-8"),
+        hashlib.sha256
+    ).hexdigest()
+
+    # Compare signatures
+    return hmac.compare_digest(expected_signature, received_signature)
+```
+
+## Event Payloads
+
+### `transcript.completed` Event
+
+This event includes a convenient URL for accessing the transcript:
+- `frontend_url`: Direct link to view the transcript in the web interface
+
+```json
+{
+  "event": "transcript.completed",
+  "event_id": "transcript.completed-abc-123-def-456",
+  "timestamp": "2025-08-27T12:34:56.789012Z",
+  "transcript": {
+    "id": "abc-123-def-456",
+    "room_id": "room-789",
+    "created_at": "2025-08-27T12:00:00Z",
+    "duration": 1800.5,
+    "title": "Q3 Product Planning Meeting",
+    "short_summary": "Team discussed Q3 product roadmap, prioritizing mobile app features and API improvements.",
+    "long_summary": "The product team met to finalize the Q3 roadmap. Key decisions included...",
+    "webvtt": "WEBVTT\n\n00:00:00.000 --> 00:00:05.000\n<v Speaker 1>Welcome everyone to today's meeting...",
+    "topics": [
+      {
+        "title": "Introduction and Agenda",
+        "summary": "Meeting kickoff with agenda review",
+        "timestamp": 0.0,
+        "duration": 120.0,
+        "webvtt": "WEBVTT\n\n00:00:00.000 --> 00:00:05.000\n<v Speaker 1>Welcome everyone..."
+      },
+      {
+        "title": "Mobile App Features Discussion",
+        "summary": "Team reviewed proposed mobile app features for Q3",
+        "timestamp": 120.0,
+        "duration": 600.0,
+        "webvtt": "WEBVTT\n\n00:02:00.000 --> 00:02:10.000\n<v Speaker 2>Let's talk about the mobile app..."
+      }
+    ],
+    "participants": [
+      {
+        "id": "participant-1",
+        "name": "John Doe",
+        "speaker": "Speaker 1"
+      },
+      {
+        "id": "participant-2",
+        "name": "Jane Smith",
+        "speaker": "Speaker 2"
+      }
+    ],
+    "source_language": "en",
+    "target_language": "en",
+    "status": "completed",
+    "frontend_url": "https://app.reflector.com/transcripts/abc-123-def-456"
+  },
+  "room": {
+    "id": "room-789",
+    "name": "Product Team Room"
+  }
+}
+```
+
+### `test` Event
+
+```json
+{
+  "event": "test",
+  "event_id": "test.2025-08-27T12:34:56.789012Z",
+  "timestamp": "2025-08-27T12:34:56.789012Z",
+  "message": "This is a test webhook from Reflector",
+  "room": {
+    "id": "room-789",
+    "name": "Product Team Room"
+  }
+}
+```
+
+## Retry Policy
+
+Webhooks are delivered with automatic retry logic to handle transient failures. When a webhook delivery fails due to server errors or network issues, Reflector will automatically retry the delivery multiple times over an extended period.
+
+### Retry Mechanism
+
+Reflector implements an exponential backoff strategy for webhook retries:
+
+- **Initial retry delay**: 60 seconds after the first failure
+- **Exponential backoff**: Each subsequent retry waits approximately twice as long as the previous one
+- **Maximum retry interval**: 1 hour (backoff is capped at this duration)
+- **Maximum retry attempts**: 30 attempts total
+- **Total retry duration**: Retries continue for approximately 24 hours
+
+### How Retries Work
+
+When a webhook fails, Reflector will:
+1. Wait 60 seconds, then retry (attempt #1)
+2. If it fails again, wait ~2 minutes, then retry (attempt #2)
+3. Continue doubling the wait time up to a maximum of 1 hour between attempts
+4. Keep retrying at 1-hour intervals until successful or 30 attempts are exhausted
+
+The `X-Webhook-Retry` header indicates the current retry attempt number (0 for the initial attempt, 1 for first retry, etc.), allowing your endpoint to track retry attempts.
+
+### Retry Behavior by HTTP Status Code
+
+| Status Code | Behavior |
+|-------------|----------|
+| 2xx (Success) | No retry, webhook marked as delivered |
+| 4xx (Client Error) | No retry, request is considered permanently failed |
+| 5xx (Server Error) | Automatic retry with exponential backoff |
+| Network/Timeout Error | Automatic retry with exponential backoff |
+
+**Important Notes:**
+- Webhooks timeout after 30 seconds. If your endpoint takes longer to respond, it will be considered a timeout error and retried.
+- During the retry period (~24 hours), you may receive the same webhook multiple times if your endpoint experiences intermittent failures.
+- There is no mechanism to manually retry failed webhooks after the retry period expires.
+
+## Testing Webhooks
+
+You can test your webhook configuration before processing transcripts:
+
+```http
+POST /v1/rooms/{room_id}/webhook/test
+```
+
+Response:
+```json
+{
+  "success": true,
+  "status_code": 200,
+  "message": "Webhook test successful",
+  "response_preview": "OK"
+}
+```
+
+Or in case of failure:
+```json
+{
+  "success": false,
+  "error": "Webhook request timed out (10 seconds)"
+}
+```

server/gpu/modal_deployments/reflector_transcriber.py

@@ -1,41 +1,78 @@
 import os
-import tempfile
+import sys
 import threading
+import uuid
+from typing import Generator, Mapping, NamedTuple, NewType, TypedDict
+from urllib.parse import urlparse
 
 import modal
-from pydantic import BaseModel
-
-MODELS_DIR = "/models"
 
 MODEL_NAME = "large-v2"
 MODEL_COMPUTE_TYPE: str = "float16"
 MODEL_NUM_WORKERS: int = 1
-
 MINUTES = 60  # seconds
+SAMPLERATE = 16000
+UPLOADS_PATH = "/uploads"
+CACHE_PATH = "/models"
+SUPPORTED_FILE_EXTENSIONS = ["mp3", "mp4", "mpeg", "mpga", "m4a", "wav", "webm"]
+VAD_CONFIG = {
+    "batch_max_duration": 30.0,
+    "silence_padding": 0.5,
+    "window_size": 512,
+}
 
-volume = modal.Volume.from_name("models", create_if_missing=True)
+
+WhisperUniqFilename = NewType("WhisperUniqFilename", str)
+AudioFileExtension = NewType("AudioFileExtension", str)
 
 app = modal.App("reflector-transcriber")
 
+model_cache = modal.Volume.from_name("models", create_if_missing=True)
+upload_volume = modal.Volume.from_name("whisper-uploads", create_if_missing=True)
+
+
+class TimeSegment(NamedTuple):
+    """Represents a time segment with start and end times."""
+
+    start: float
+    end: float
+
+
+class AudioSegment(NamedTuple):
+    """Represents an audio segment with timing and audio data."""
+
+    start: float
+    end: float
+    audio: any
+
+
+class TranscriptResult(NamedTuple):
+    """Represents a transcription result with text and word timings."""
+
+    text: str
+    words: list["WordTiming"]
+
+
+class WordTiming(TypedDict):
+    """Represents a word with its timing information."""
+
+    word: str
+    start: float
+    end: float
+
 
 def download_model():
     from faster_whisper import download_model
 
-    volume.reload()
+    model_cache.reload()
 
-    download_model(MODEL_NAME, cache_dir=MODELS_DIR)
+    download_model(MODEL_NAME, cache_dir=CACHE_PATH)
 
-    volume.commit()
+    model_cache.commit()
 
 
 image = (
     modal.Image.debian_slim(python_version="3.12")
-    .pip_install(
-        "huggingface_hub==0.27.1",
-        "hf-transfer==0.1.9",
-        "torch==2.5.1",
-        "faster-whisper==1.1.1",
-    )
     .env(
         {
             "HF_HUB_ENABLE_HF_TRANSFER": "1",
@@ -45,19 +82,98 @@ image = (
             ),
         }
     )
-    .run_function(download_model, volumes={MODELS_DIR: volume})
+    .apt_install("ffmpeg")
+    .pip_install(
+        "huggingface_hub==0.27.1",
+        "hf-transfer==0.1.9",
+        "torch==2.5.1",
+        "faster-whisper==1.1.1",
+        "fastapi==0.115.12",
+        "requests",
+        "librosa==0.10.1",
+        "numpy<2",
+        "silero-vad==5.1.0",
+    )
+    .run_function(download_model, volumes={CACHE_PATH: model_cache})
 )
 
 
+def detect_audio_format(url: str, headers: Mapping[str, str]) -> AudioFileExtension:
+    parsed_url = urlparse(url)
+    url_path = parsed_url.path
+
+    for ext in SUPPORTED_FILE_EXTENSIONS:
+        if url_path.lower().endswith(f".{ext}"):
+            return AudioFileExtension(ext)
+
+    content_type = headers.get("content-type", "").lower()
+    if "audio/mpeg" in content_type or "audio/mp3" in content_type:
+        return AudioFileExtension("mp3")
+    if "audio/wav" in content_type:
+        return AudioFileExtension("wav")
+    if "audio/mp4" in content_type:
+        return AudioFileExtension("mp4")
+
+    raise ValueError(
+        f"Unsupported audio format for URL: {url}. "
+        f"Supported extensions: {', '.join(SUPPORTED_FILE_EXTENSIONS)}"
+    )
+
+
+def download_audio_to_volume(
+    audio_file_url: str,
+) -> tuple[WhisperUniqFilename, AudioFileExtension]:
+    import requests
+    from fastapi import HTTPException
+
+    response = requests.head(audio_file_url, allow_redirects=True)
+    if response.status_code == 404:
+        raise HTTPException(status_code=404, detail="Audio file not found")
+
+    response = requests.get(audio_file_url, allow_redirects=True)
+    response.raise_for_status()
+
+    audio_suffix = detect_audio_format(audio_file_url, response.headers)
+    unique_filename = WhisperUniqFilename(f"{uuid.uuid4()}.{audio_suffix}")
+    file_path = f"{UPLOADS_PATH}/{unique_filename}"
+
+    with open(file_path, "wb") as f:
+        f.write(response.content)
+
+    upload_volume.commit()
+    return unique_filename, audio_suffix
+
+
+def pad_audio(audio_array, sample_rate: int = SAMPLERATE):
+    """Add 0.5s of silence if audio is shorter than the silence_padding window.
+
+    Whisper does not require this strictly, but aligning behavior with Parakeet
+    avoids edge-case crashes on extremely short inputs and makes comparisons easier.
+    """
+    import numpy as np
+
+    audio_duration = len(audio_array) / sample_rate
+    if audio_duration < VAD_CONFIG["silence_padding"]:
+        silence_samples = int(sample_rate * VAD_CONFIG["silence_padding"])
+        silence = np.zeros(silence_samples, dtype=np.float32)
+        return np.concatenate([audio_array, silence])
+    return audio_array
+
+
 @app.cls(
     gpu="A10G",
     timeout=5 * MINUTES,
     scaledown_window=5 * MINUTES,
-    allow_concurrent_inputs=6,
     image=image,
-    volumes={MODELS_DIR: volume},
+    volumes={CACHE_PATH: model_cache, UPLOADS_PATH: upload_volume},
 )
-class Transcriber:
+@modal.concurrent(max_inputs=10)
+class TranscriberWhisperLive:
+    """Live transcriber class for small audio segments (A10G).
+
+    Mirrors the Parakeet live class API but uses Faster-Whisper under the hood.
+    """
+
     @modal.enter()
     def enter(self):
         import faster_whisper
@@ -71,23 +187,200 @@ class Transcriber:
             device=self.device,
             compute_type=MODEL_COMPUTE_TYPE,
             num_workers=MODEL_NUM_WORKERS,
-            download_root=MODELS_DIR,
+            download_root=CACHE_PATH,
             local_files_only=True,
         )
+        print(f"Model is on device: {self.device}")
 
     @modal.method()
     def transcribe_segment(
         self,
-        audio_data: str,
-        audio_suffix: str,
-        language: str,
+        filename: str,
+        language: str = "en",
     ):
-        with tempfile.NamedTemporaryFile("wb+", suffix=f".{audio_suffix}") as fp:
-            fp.write(audio_data)
+        """Transcribe a single uploaded audio file by filename."""
+        upload_volume.reload()
+
+        file_path = f"{UPLOADS_PATH}/{filename}"
+        if not os.path.exists(file_path):
+            raise FileNotFoundError(f"File not found: {file_path}")
+
+        with self.lock:
+            with NoStdStreams():
+                segments, _ = self.model.transcribe(
+                    file_path,
+                    language=language,
+                    beam_size=5,
+                    word_timestamps=True,
+                    vad_filter=True,
+                    vad_parameters={"min_silence_duration_ms": 500},
+                )
+
+        segments = list(segments)
+        text = "".join(segment.text for segment in segments).strip()
+        words = [
+            {
+                "word": word.word,
+                "start": round(float(word.start), 2),
+                "end": round(float(word.end), 2),
+            }
+            for segment in segments
+            for word in segment.words
+        ]
+
+        return {"text": text, "words": words}
+
+    @modal.method()
+    def transcribe_batch(
+        self,
+        filenames: list[str],
+        language: str = "en",
+    ):
+        """Transcribe multiple uploaded audio files and return per-file results."""
+        upload_volume.reload()
+
+        results = []
+        for filename in filenames:
+            file_path = f"{UPLOADS_PATH}/{filename}"
+            if not os.path.exists(file_path):
+                raise FileNotFoundError(f"Batch file not found: {file_path}")
+
+            with self.lock:
+                with NoStdStreams():
+                    segments, _ = self.model.transcribe(
+                        file_path,
+                        language=language,
+                        beam_size=5,
+                        word_timestamps=True,
+                        vad_filter=True,
+                        vad_parameters={"min_silence_duration_ms": 500},
+                    )
+
+            segments = list(segments)
+            text = "".join(seg.text for seg in segments).strip()
+            words = [
+                {
+                    "word": w.word,
+                    "start": round(float(w.start), 2),
+                    "end": round(float(w.end), 2),
+                }
+                for seg in segments
+                for w in seg.words
+            ]
+
+            results.append(
+                {
+                    "filename": filename,
+                    "text": text,
+                    "words": words,
+                }
+            )
+
+        return results
+
+
+@app.cls(
+    gpu="L40S",
+    timeout=15 * MINUTES,
+    image=image,
+    volumes={CACHE_PATH: model_cache, UPLOADS_PATH: upload_volume},
+)
+class TranscriberWhisperFile:
+    """File transcriber for larger/longer audio, using VAD-driven batching (L40S)."""
+
+    @modal.enter()
+    def enter(self):
+        import faster_whisper
+        import torch
+        from silero_vad import load_silero_vad
+
+        self.lock = threading.Lock()
+        self.use_gpu = torch.cuda.is_available()
+        self.device = "cuda" if self.use_gpu else "cpu"
+        self.model = faster_whisper.WhisperModel(
+            MODEL_NAME,
+            device=self.device,
+            compute_type=MODEL_COMPUTE_TYPE,
+            num_workers=MODEL_NUM_WORKERS,
+            download_root=CACHE_PATH,
+            local_files_only=True,
+        )
+        self.vad_model = load_silero_vad(onnx=False)
+
+    @modal.method()
+    def transcribe_segment(
+        self, filename: str, timestamp_offset: float = 0.0, language: str = "en"
+    ):
+        import librosa
+        import numpy as np
+        from silero_vad import VADIterator
+
+        def vad_segments(
+            audio_array,
+            sample_rate: int = SAMPLERATE,
+            window_size: int = VAD_CONFIG["window_size"],
+        ) -> Generator[TimeSegment, None, None]:
+            """Generate speech segments as TimeSegment using Silero VAD."""
+            iterator = VADIterator(self.vad_model, sampling_rate=sample_rate)
+            start = None
+            for i in range(0, len(audio_array), window_size):
+                chunk = audio_array[i : i + window_size]
+                if len(chunk) < window_size:
+                    chunk = np.pad(
+                        chunk, (0, window_size - len(chunk)), mode="constant"
+                    )
+                speech = iterator(chunk)
+                if not speech:
+                    continue
+                if "start" in speech:
+                    start = speech["start"]
+                    continue
+                if "end" in speech and start is not None:
+                    end = speech["end"]
+                    yield TimeSegment(
+                        start / float(SAMPLERATE), end / float(SAMPLERATE)
+                    )
+                    start = None
+            iterator.reset_states()
+
+        upload_volume.reload()
+        file_path = f"{UPLOADS_PATH}/{filename}"
+        if not os.path.exists(file_path):
+            raise FileNotFoundError(f"File not found: {file_path}")
+
+        audio_array, _sr = librosa.load(file_path, sr=SAMPLERATE, mono=True)
+
+        # Batch segments up to ~30s windows by merging contiguous VAD segments
+        merged_batches: list[TimeSegment] = []
+        batch_start = None
+        batch_end = None
+        max_duration = VAD_CONFIG["batch_max_duration"]
+        for segment in vad_segments(audio_array):
+            seg_start, seg_end = segment.start, segment.end
+            if batch_start is None:
+                batch_start, batch_end = seg_start, seg_end
+                continue
+            if seg_end - batch_start <= max_duration:
+                batch_end = seg_end
+            else:
+                merged_batches.append(TimeSegment(batch_start, batch_end))
+                batch_start, batch_end = seg_start, seg_end
+        if batch_start is not None and batch_end is not None:
+            merged_batches.append(TimeSegment(batch_start, batch_end))
+
+        all_text = []
+        all_words = []
+
+        for segment in merged_batches:
+            start_time, end_time = segment.start, segment.end
+            s_idx = int(start_time * SAMPLERATE)
+            e_idx = int(end_time * SAMPLERATE)
+            segment = audio_array[s_idx:e_idx]
+            segment = pad_audio(segment, SAMPLERATE)
 
             with self.lock:
                 segments, _ = self.model.transcribe(
-                    fp.name,
+                    segment,
                     language=language,
                     beam_size=5,
                     word_timestamps=True,
@@ -96,66 +389,220 @@ class Transcriber:
                 )
 
             segments = list(segments)
-            text = "".join(segment.text for segment in segments)
+            text = "".join(seg.text for seg in segments).strip()
             words = [
-                {"word": word.word, "start": word.start, "end": word.end}
-                for segment in segments
-                for word in segment.words
+                {
+                    "word": w.word,
+                    "start": round(float(w.start) + start_time + timestamp_offset, 2),
+                    "end": round(float(w.end) + start_time + timestamp_offset, 2),
+                }
+                for seg in segments
+                for w in seg.words
             ]
+            if text:
+                all_text.append(text)
+            all_words.extend(words)
 
-            return {"text": text, "words": words}
+        return {"text": " ".join(all_text), "words": all_words}
+
+
+def detect_audio_format(url: str, headers: dict) -> str:
+    from urllib.parse import urlparse
+
+    from fastapi import HTTPException
+
+    url_path = urlparse(url).path
+    for ext in SUPPORTED_FILE_EXTENSIONS:
+        if url_path.lower().endswith(f".{ext}"):
+            return ext
+
+    content_type = headers.get("content-type", "").lower()
+    if "audio/mpeg" in content_type or "audio/mp3" in content_type:
+        return "mp3"
+    if "audio/wav" in content_type:
+        return "wav"
+    if "audio/mp4" in content_type:
+        return "mp4"
+
+    raise HTTPException(
+        status_code=400,
+        detail=(
+            f"Unsupported audio format for URL. Supported extensions: {', '.join(SUPPORTED_FILE_EXTENSIONS)}"
+        ),
+    )
+
+
+def download_audio_to_volume(audio_file_url: str) -> tuple[str, str]:
+    import requests
+    from fastapi import HTTPException
+
+    response = requests.head(audio_file_url, allow_redirects=True)
+    if response.status_code == 404:
+        raise HTTPException(status_code=404, detail="Audio file not found")
+
+    response = requests.get(audio_file_url, allow_redirects=True)
+    response.raise_for_status()
+
+    audio_suffix = detect_audio_format(audio_file_url, response.headers)
+    unique_filename = f"{uuid.uuid4()}.{audio_suffix}"
+    file_path = f"{UPLOADS_PATH}/{unique_filename}"
+
+    with open(file_path, "wb") as f:
+        f.write(response.content)
+
+    upload_volume.commit()
+    return unique_filename, audio_suffix
 
 
 @app.function(
     scaledown_window=60,
-    timeout=60,
-    allow_concurrent_inputs=40,
+    timeout=600,
     secrets=[
         modal.Secret.from_name("reflector-gpu"),
     ],
-    volumes={MODELS_DIR: volume},
+    volumes={CACHE_PATH: model_cache, UPLOADS_PATH: upload_volume},
+    image=image,
 )
+@modal.concurrent(max_inputs=40)
 @modal.asgi_app()
 def web():
-    from fastapi import Body, Depends, FastAPI, HTTPException, UploadFile, status
+    from fastapi import (
+        Body,
+        Depends,
+        FastAPI,
+        Form,
+        HTTPException,
+        UploadFile,
+        status,
+    )
     from fastapi.security import OAuth2PasswordBearer
-    from typing_extensions import Annotated
 
-    transcriber = Transcriber()
+    transcriber_live = TranscriberWhisperLive()
+    transcriber_file = TranscriberWhisperFile()
 
     app = FastAPI()
 
     oauth2_scheme = OAuth2PasswordBearer(tokenUrl="token")
 
-    supported_file_types = ["mp3", "mp4", "mpeg", "mpga", "m4a", "wav", "webm"]
-
     def apikey_auth(apikey: str = Depends(oauth2_scheme)):
-        if apikey != os.environ["REFLECTOR_GPU_APIKEY"]:
-            raise HTTPException(
-                status_code=status.HTTP_401_UNAUTHORIZED,
-                detail="Invalid API key",
-                headers={"WWW-Authenticate": "Bearer"},
-            )
+        if apikey == os.environ["REFLECTOR_GPU_APIKEY"]:
+            return
+        raise HTTPException(
+            status_code=status.HTTP_401_UNAUTHORIZED,
+            detail="Invalid API key",
+            headers={"WWW-Authenticate": "Bearer"},
+        )
 
-    class TranscriptResponse(BaseModel):
-        result: dict
+    class TranscriptResponse(dict):
+        pass
 
     @app.post("/v1/audio/transcriptions", dependencies=[Depends(apikey_auth)])
     def transcribe(
-        file: UploadFile,
-        model: str = "whisper-1",
-        language: Annotated[str, Body(...)] = "en",
-    ) -> TranscriptResponse:
-        audio_data = file.file.read()
-        audio_suffix = file.filename.split(".")[-1]
-        assert audio_suffix in supported_file_types
+        file: UploadFile = None,
+        files: list[UploadFile] | None = None,
+        model: str = Form(MODEL_NAME),
+        language: str = Form("en"),
+        batch: bool = Form(False),
+    ):
+        if not file and not files:
+            raise HTTPException(
+                status_code=400, detail="Either 'file' or 'files' parameter is required"
+            )
+        if batch and not files:
+            raise HTTPException(
+                status_code=400, detail="Batch transcription requires 'files'"
+            )
 
-        func = transcriber.transcribe_segment.spawn(
-            audio_data=audio_data,
-            audio_suffix=audio_suffix,
-            language=language,
-        )
-        result = func.get()
-        return result
+        upload_files = [file] if file else files
+
+        uploaded_filenames: list[str] = []
+        for upload_file in upload_files:
+            audio_suffix = upload_file.filename.split(".")[-1]
+            if audio_suffix not in SUPPORTED_FILE_EXTENSIONS:
+                raise HTTPException(
+                    status_code=400,
+                    detail=(
+                        f"Unsupported audio format. Supported extensions: {', '.join(SUPPORTED_FILE_EXTENSIONS)}"
+                    ),
+                )
+
+            unique_filename = f"{uuid.uuid4()}.{audio_suffix}"
+            file_path = f"{UPLOADS_PATH}/{unique_filename}"
+            with open(file_path, "wb") as f:
+                content = upload_file.file.read()
+                f.write(content)
+            uploaded_filenames.append(unique_filename)
+
+        upload_volume.commit()
+
+        try:
+            if batch and len(upload_files) > 1:
+                func = transcriber_live.transcribe_batch.spawn(
+                    filenames=uploaded_filenames,
+                    language=language,
+                )
+                results = func.get()
+                return {"results": results}
+
+            results = []
+            for filename in uploaded_filenames:
+                func = transcriber_live.transcribe_segment.spawn(
+                    filename=filename,
+                    language=language,
+                )
+                result = func.get()
+                result["filename"] = filename
+                results.append(result)
+
+            return {"results": results} if len(results) > 1 else results[0]
+        finally:
+            for filename in uploaded_filenames:
+                try:
+                    file_path = f"{UPLOADS_PATH}/{filename}"
+                    os.remove(file_path)
+                except Exception:
+                    pass
+            upload_volume.commit()
+
+    @app.post("/v1/audio/transcriptions-from-url", dependencies=[Depends(apikey_auth)])
+    def transcribe_from_url(
+        audio_file_url: str = Body(
+            ..., description="URL of the audio file to transcribe"
+        ),
+        model: str = Body(MODEL_NAME),
+        language: str = Body("en"),
+        timestamp_offset: float = Body(0.0),
+    ):
+        unique_filename, _audio_suffix = download_audio_to_volume(audio_file_url)
+        try:
+            func = transcriber_file.transcribe_segment.spawn(
+                filename=unique_filename,
+                timestamp_offset=timestamp_offset,
+                language=language,
+            )
+            result = func.get()
+            return result
+        finally:
+            try:
+                file_path = f"{UPLOADS_PATH}/{unique_filename}"
+                os.remove(file_path)
+                upload_volume.commit()
+            except Exception:
+                pass
 
     return app
+
+
+class NoStdStreams:
+    def __init__(self):
+        self.devnull = open(os.devnull, "w")
+
+    def __enter__(self):
+        self._stdout, self._stderr = sys.stdout, sys.stderr
+        self._stdout.flush()
+        self._stderr.flush()
+        sys.stdout, sys.stderr = self.devnull, self.devnull
+
+    def __exit__(self, exc_type, exc_value, traceback):
+        sys.stdout, sys.stderr = self._stdout, self._stderr
+        self.devnull.close()

server/gpu/modal_deployments/reflector_transcriber_parakeet.py

@@ -3,7 +3,7 @@ import os
 import sys
 import threading
 import uuid
-from typing import Mapping, NewType
+from typing import Generator, Mapping, NamedTuple, NewType, TypedDict
 from urllib.parse import urlparse
 
 import modal
@@ -14,10 +14,7 @@ SAMPLERATE = 16000
 UPLOADS_PATH = "/uploads"
 CACHE_PATH = "/cache"
 VAD_CONFIG = {
-    "max_segment_duration": 30.0,
-    "batch_max_files": 10,
-    "batch_max_duration": 5.0,
-    "min_segment_duration": 0.02,
+    "batch_max_duration": 30.0,
     "silence_padding": 0.5,
     "window_size": 512,
 }
@@ -25,6 +22,37 @@ VAD_CONFIG = {
 ParakeetUniqFilename = NewType("ParakeetUniqFilename", str)
 AudioFileExtension = NewType("AudioFileExtension", str)
 
+
+class TimeSegment(NamedTuple):
+    """Represents a time segment with start and end times."""
+
+    start: float
+    end: float
+
+
+class AudioSegment(NamedTuple):
+    """Represents an audio segment with timing and audio data."""
+
+    start: float
+    end: float
+    audio: any
+
+
+class TranscriptResult(NamedTuple):
+    """Represents a transcription result with text and word timings."""
+
+    text: str
+    words: list["WordTiming"]
+
+
+class WordTiming(TypedDict):
+    """Represents a word with its timing information."""
+
+    word: str
+    start: float
+    end: float
+
+
 app = modal.App("reflector-transcriber-parakeet")
 
 # Volume for caching model weights
@@ -170,12 +198,14 @@ class TranscriberParakeetLive:
                 (output,) = self.model.transcribe([padded_audio], timestamps=True)
 
         text = output.text.strip()
-        words = [
-            {
-                "word": word_info["word"],
-                "start": round(word_info["start"], 2),
-                "end": round(word_info["end"], 2),
-            }
+        words: list[WordTiming] = [
+            WordTiming(
+                # XXX the space added here is to match the output of whisper
+                # whisper add space to each words, while parakeet don't
+                word=word_info["word"] + " ",
+                start=round(word_info["start"], 2),
+                end=round(word_info["end"], 2),
+            )
             for word_info in output.timestamp["word"]
         ]
 
@@ -211,12 +241,12 @@ class TranscriberParakeetLive:
         for i, (filename, output) in enumerate(zip(filenames, outputs)):
             text = output.text.strip()
 
-            words = [
-                {
-                    "word": word_info["word"],
-                    "start": round(word_info["start"], 2),
-                    "end": round(word_info["end"], 2),
-                }
+            words: list[WordTiming] = [
+                WordTiming(
+                    word=word_info["word"] + " ",
+                    start=round(word_info["start"], 2),
+                    end=round(word_info["end"], 2),
+                )
                 for word_info in output.timestamp["word"]
             ]
 
@@ -271,7 +301,9 @@ class TranscriberParakeetFile:
             audio_array, sample_rate = librosa.load(file_path, sr=SAMPLERATE, mono=True)
             return audio_array
 
-        def vad_segment_generator(audio_array):
+        def vad_segment_generator(
+            audio_array,
+        ) -> Generator[TimeSegment, None, None]:
             """Generate speech segments using VAD with start/end sample indices"""
             vad_iterator = VADIterator(self.vad_model, sampling_rate=SAMPLERATE)
             window_size = VAD_CONFIG["window_size"]
@@ -297,107 +329,121 @@ class TranscriberParakeetFile:
                     start_time = start / float(SAMPLERATE)
                     end_time = end / float(SAMPLERATE)
 
-                    # Extract the actual audio segment
-                    audio_segment = audio_array[start:end]
-
-                    yield (start_time, end_time, audio_segment)
+                    yield TimeSegment(start_time, end_time)
                     start = None
 
             vad_iterator.reset_states()
 
-        def vad_segment_filter(segments):
-            """Filter VAD segments by duration and chunk large segments"""
-            min_dur = VAD_CONFIG["min_segment_duration"]
-            max_dur = VAD_CONFIG["max_segment_duration"]
+        def batch_speech_segments(
+            segments: Generator[TimeSegment, None, None], max_duration: int
+        ) -> Generator[TimeSegment, None, None]:
+            """
+            Input segments:
+              [0-2] [3-5] [6-8] [10-11] [12-15] [17-19] [20-22]
 
-            for start_time, end_time, audio_segment in segments:
-                segment_duration = end_time - start_time
+                                  ↓ (max_duration=10)
 
-                # Skip very small segments
-                if segment_duration < min_dur:
+              Output batches:
+              [0-8]           [10-19]          [20-22]
+
+            Note: silences are kept for better transcription, previous implementation was
+            passing segments separatly, but the output was less accurate.
+            """
+            batch_start_time = None
+            batch_end_time = None
+
+            for segment in segments:
+                start_time, end_time = segment.start, segment.end
+                if batch_start_time is None or batch_end_time is None:
+                    batch_start_time = start_time
+                    batch_end_time = end_time
                     continue
 
-                # If segment is within max duration, yield as-is
-                if segment_duration <= max_dur:
-                    yield (start_time, end_time, audio_segment)
+                total_duration = end_time - batch_start_time
+
+                if total_duration <= max_duration:
+                    batch_end_time = end_time
                     continue
 
-                # Chunk large segments into smaller pieces
-                chunk_samples = int(max_dur * SAMPLERATE)
-                current_start = start_time
+                yield TimeSegment(batch_start_time, batch_end_time)
+                batch_start_time = start_time
+                batch_end_time = end_time
 
-                for chunk_offset in range(0, len(audio_segment), chunk_samples):
-                    chunk_audio = audio_segment[
-                        chunk_offset : chunk_offset + chunk_samples
-                    ]
-                    if len(chunk_audio) == 0:
-                        break
+            if batch_start_time is None or batch_end_time is None:
+                return
 
-                    chunk_duration = len(chunk_audio) / float(SAMPLERATE)
-                    chunk_end = current_start + chunk_duration
+            yield TimeSegment(batch_start_time, batch_end_time)
 
-                    # Only yield chunks that meet minimum duration
-                    if chunk_duration >= min_dur:
-                        yield (current_start, chunk_end, chunk_audio)
+        def batch_segment_to_audio_segment(
+            segments: Generator[TimeSegment, None, None],
+            audio_array,
+        ) -> Generator[AudioSegment, None, None]:
+            """Extract audio segments and apply padding for Parakeet compatibility.
 
-                    current_start = chunk_end
+            Uses pad_audio to ensure segments are at least 0.5s long, preventing
+            Parakeet crashes. This padding may cause slight timing overlaps between
+            segments, which are corrected by enforce_word_timing_constraints.
+            """
+            for segment in segments:
+                start_time, end_time = segment.start, segment.end
+                start_sample = int(start_time * SAMPLERATE)
+                end_sample = int(end_time * SAMPLERATE)
+                audio_segment = audio_array[start_sample:end_sample]
 
-        def batch_segments(segments, max_files=10, max_duration=5.0):
-            batch = []
-            batch_duration = 0.0
+                padded_segment = pad_audio(audio_segment, SAMPLERATE)
 
-            for start_time, end_time, audio_segment in segments:
-                segment_duration = end_time - start_time
+                yield AudioSegment(start_time, end_time, padded_segment)
 
-                if segment_duration < VAD_CONFIG["silence_padding"]:
-                    silence_samples = int(
-                        (VAD_CONFIG["silence_padding"] - segment_duration) * SAMPLERATE
-                    )
-                    padding = np.zeros(silence_samples, dtype=np.float32)
-                    audio_segment = np.concatenate([audio_segment, padding])
-                    segment_duration = VAD_CONFIG["silence_padding"]
-
-                batch.append((start_time, end_time, audio_segment))
-                batch_duration += segment_duration
-
-                if len(batch) >= max_files or batch_duration >= max_duration:
-                    yield batch
-                    batch = []
-                    batch_duration = 0.0
-
-            if batch:
-                yield batch
-
-        def transcribe_batch(model, audio_segments):
+        def transcribe_batch(model, audio_segments: list) -> list:
             with NoStdStreams():
                 outputs = model.transcribe(audio_segments, timestamps=True)
             return outputs
 
+        def enforce_word_timing_constraints(
+            words: list[WordTiming],
+        ) -> list[WordTiming]:
+            """Enforce that word end times don't exceed the start time of the next word.
+
+            Due to silence padding added in batch_segment_to_audio_segment for better
+            transcription accuracy, word timings from different segments may overlap.
+            This function ensures there are no overlaps by adjusting end times.
+            """
+            if len(words) <= 1:
+                return words
+
+            enforced_words = []
+            for i, word in enumerate(words):
+                enforced_word = word.copy()
+
+                if i < len(words) - 1:
+                    next_start = words[i + 1]["start"]
+                    if enforced_word["end"] > next_start:
+                        enforced_word["end"] = next_start
+
+                enforced_words.append(enforced_word)
+
+            return enforced_words
+
         def emit_results(
-            results,
-            segments_info,
-            batch_index,
-            total_batches,
-        ):
+            results: list,
+            segments_info: list[AudioSegment],
+        ) -> Generator[TranscriptResult, None, None]:
             """Yield transcribed text and word timings from model output, adjusting timestamps to absolute positions."""
-            for i, (output, (start_time, end_time, _)) in enumerate(
-                zip(results, segments_info)
-            ):
+            for i, (output, segment) in enumerate(zip(results, segments_info)):
+                start_time, end_time = segment.start, segment.end
                 text = output.text.strip()
-                words = [
-                    {
-                        "word": word_info["word"],
-                        "start": round(
+                words: list[WordTiming] = [
+                    WordTiming(
+                        word=word_info["word"] + " ",
+                        start=round(
                             word_info["start"] + start_time + timestamp_offset, 2
                         ),
-                        "end": round(
-                            word_info["end"] + start_time + timestamp_offset, 2
-                        ),
-                    }
+                        end=round(word_info["end"] + start_time + timestamp_offset, 2),
+                    )
                     for word_info in output.timestamp["word"]
                 ]
 
-                yield text, words
+                yield TranscriptResult(text, words)
 
         upload_volume.reload()
 
@@ -407,41 +453,31 @@ class TranscriberParakeetFile:
 
         audio_array = load_and_convert_audio(file_path)
         total_duration = len(audio_array) / float(SAMPLERATE)
-        processed_duration = 0.0
 
-        all_text_parts = []
-        all_words = []
+        all_text_parts: list[str] = []
+        all_words: list[WordTiming] = []
 
         raw_segments = vad_segment_generator(audio_array)
-        filtered_segments = vad_segment_filter(raw_segments)
-        batches = batch_segments(
-            filtered_segments,
-            VAD_CONFIG["batch_max_files"],
+        speech_segments = batch_speech_segments(
+            raw_segments,
             VAD_CONFIG["batch_max_duration"],
         )
+        audio_segments = batch_segment_to_audio_segment(speech_segments, audio_array)
 
-        batch_index = 0
-        total_batches = max(
-            1, int(total_duration / VAD_CONFIG["batch_max_duration"]) + 1
-        )
+        for batch in audio_segments:
+            audio_segment = batch.audio
+            results = transcribe_batch(self.model, [audio_segment])
 
-        for batch in batches:
-            batch_index += 1
-            audio_segments = [seg[2] for seg in batch]
-            results = transcribe_batch(self.model, audio_segments)
-
-            for text, words in emit_results(
+            for result in emit_results(
                 results,
-                batch,
-                batch_index,
-                total_batches,
+                [batch],
             ):
-                if not text:
+                if not result.text:
                     continue
-                all_text_parts.append(text)
-                all_words.extend(words)
+                all_text_parts.append(result.text)
+                all_words.extend(result.words)
 
-            processed_duration += sum(len(seg[2]) / float(SAMPLERATE) for seg in batch)
+        all_words = enforce_word_timing_constraints(all_words)
 
         combined_text = " ".join(all_text_parts)
         return {"text": combined_text, "words": all_words}

server/migrations/versions/0194f65cd6d3_add_webhook_fields_to_rooms.py

@@ -0,0 +1,36 @@
+"""Add webhook fields to rooms
+
+Revision ID: 0194f65cd6d3
+Revises: 5a8907fd1d78
+Create Date: 2025-08-27 09:03:19.610995
+
+"""
+
+from typing import Sequence, Union
+
+import sqlalchemy as sa
+from alembic import op
+
+# revision identifiers, used by Alembic.
+revision: str = "0194f65cd6d3"
+down_revision: Union[str, None] = "5a8907fd1d78"
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+
+def upgrade() -> None:
+    # ### commands auto generated by Alembic - please adjust! ###
+    with op.batch_alter_table("room", schema=None) as batch_op:
+        batch_op.add_column(sa.Column("webhook_url", sa.String(), nullable=True))
+        batch_op.add_column(sa.Column("webhook_secret", sa.String(), nullable=True))
+
+    # ### end Alembic commands ###
+
+
+def downgrade() -> None:
+    # ### commands auto generated by Alembic - please adjust! ###
+    with op.batch_alter_table("room", schema=None) as batch_op:
+        batch_op.drop_column("webhook_secret")
+        batch_op.drop_column("webhook_url")
+
+    # ### end Alembic commands ###

server/migrations/versions/0ce521cda2ee_remove_user_id_from_meeting_table.py

@@ -0,0 +1,36 @@
+"""remove user_id from meeting table
+
+Revision ID: 0ce521cda2ee
+Revises: 6dec9fb5b46c
+Create Date: 2025-09-10 12:40:55.688899
+
+"""
+
+from typing import Sequence, Union
+
+import sqlalchemy as sa
+from alembic import op
+
+# revision identifiers, used by Alembic.
+revision: str = "0ce521cda2ee"
+down_revision: Union[str, None] = "6dec9fb5b46c"
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+
+def upgrade() -> None:
+    # ### commands auto generated by Alembic - please adjust! ###
+    with op.batch_alter_table("meeting", schema=None) as batch_op:
+        batch_op.drop_column("user_id")
+
+    # ### end Alembic commands ###
+
+
+def downgrade() -> None:
+    # ### commands auto generated by Alembic - please adjust! ###
+    with op.batch_alter_table("meeting", schema=None) as batch_op:
+        batch_op.add_column(
+            sa.Column("user_id", sa.VARCHAR(), autoincrement=False, nullable=True)
+        )
+
+    # ### end Alembic commands ###

server/migrations/versions/2890b5104577_add_events_column_to_meetings_table.py

@@ -0,0 +1,38 @@
+"""Add events column to meetings table
+
+Revision ID: 2890b5104577
+Revises: 6e6ea8e607c5
+Create Date: 2025-09-02 17:51:41.620777
+
+"""
+
+from typing import Sequence, Union
+
+import sqlalchemy as sa
+from alembic import op
+
+# revision identifiers, used by Alembic.
+revision: str = "2890b5104577"
+down_revision: Union[str, None] = "6e6ea8e607c5"
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+
+def upgrade() -> None:
+    # ### commands auto generated by Alembic - please adjust! ###
+    with op.batch_alter_table("meeting", schema=None) as batch_op:
+        batch_op.add_column(
+            sa.Column(
+                "events", sa.JSON(), server_default=sa.text("'[]'"), nullable=False
+            )
+        )
+
+    # ### end Alembic commands ###
+
+
+def downgrade() -> None:
+    # ### commands auto generated by Alembic - please adjust! ###
+    with op.batch_alter_table("meeting", schema=None) as batch_op:
+        batch_op.drop_column("events")
+
+    # ### end Alembic commands ###

server/migrations/versions/2ae3db106d4e_clean_up_orphaned_room_id_references_in_.py

@@ -0,0 +1,32 @@
+"""clean up orphaned room_id references in meeting table
+
+Revision ID: 2ae3db106d4e
+Revises: def1b5867d4c
+Create Date: 2025-09-11 10:35:15.759967
+
+"""
+
+from typing import Sequence, Union
+
+from alembic import op
+
+# revision identifiers, used by Alembic.
+revision: str = "2ae3db106d4e"
+down_revision: Union[str, None] = "def1b5867d4c"
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+
+def upgrade() -> None:
+    # Set room_id to NULL for meetings that reference non-existent rooms
+    op.execute("""
+        UPDATE meeting
+        SET room_id = NULL
+        WHERE room_id IS NOT NULL
+          AND room_id NOT IN (SELECT id FROM room WHERE id IS NOT NULL)
+    """)
+
+
+def downgrade() -> None:
+    # Cannot restore orphaned references - no operation needed
+    pass

server/migrations/versions/5a8907fd1d78_add_cascade_delete_to_meeting_consent_.py

@@ -0,0 +1,50 @@
+"""add cascade delete to meeting consent foreign key
+
+Revision ID: 5a8907fd1d78
+Revises: 0ab2d7ffaa16
+Create Date: 2025-08-26 17:26:50.945491
+
+"""
+
+from typing import Sequence, Union
+
+from alembic import op
+
+# revision identifiers, used by Alembic.
+revision: str = "5a8907fd1d78"
+down_revision: Union[str, None] = "0ab2d7ffaa16"
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+
+def upgrade() -> None:
+    # ### commands auto generated by Alembic - please adjust! ###
+    with op.batch_alter_table("meeting_consent", schema=None) as batch_op:
+        batch_op.drop_constraint(
+            batch_op.f("meeting_consent_meeting_id_fkey"), type_="foreignkey"
+        )
+        batch_op.create_foreign_key(
+            batch_op.f("meeting_consent_meeting_id_fkey"),
+            "meeting",
+            ["meeting_id"],
+            ["id"],
+            ondelete="CASCADE",
+        )
+
+    # ### end Alembic commands ###
+
+
+def downgrade() -> None:
+    # ### commands auto generated by Alembic - please adjust! ###
+    with op.batch_alter_table("meeting_consent", schema=None) as batch_op:
+        batch_op.drop_constraint(
+            batch_op.f("meeting_consent_meeting_id_fkey"), type_="foreignkey"
+        )
+        batch_op.create_foreign_key(
+            batch_op.f("meeting_consent_meeting_id_fkey"),
+            "meeting",
+            ["meeting_id"],
+            ["id"],
+        )
+
+    # ### end Alembic commands ###

server/migrations/versions/61882a919591_webhook_url_and_secret_null_by_default.py

@@ -0,0 +1,28 @@
+"""webhook url and secret null by default
+
+
+Revision ID: 61882a919591
+Revises: 0194f65cd6d3
+Create Date: 2025-08-29 11:46:36.738091
+
+"""
+
+from typing import Sequence, Union
+
+# revision identifiers, used by Alembic.
+revision: str = "61882a919591"
+down_revision: Union[str, None] = "0194f65cd6d3"
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+
+def upgrade() -> None:
+    # ### commands auto generated by Alembic - please adjust! ###
+    pass
+    # ### end Alembic commands ###
+
+
+def downgrade() -> None:
+    # ### commands auto generated by Alembic - please adjust! ###
+    pass
+    # ### end Alembic commands ###

server/migrations/versions/6dec9fb5b46c_make_meeting_room_id_required_and_add_.py

@@ -0,0 +1,38 @@
+"""make meeting room_id required and add foreign key
+
+Revision ID: 6dec9fb5b46c
+Revises: 61882a919591
+Create Date: 2025-09-10 10:47:06.006819
+
+"""
+
+from typing import Sequence, Union
+
+import sqlalchemy as sa
+from alembic import op
+
+# revision identifiers, used by Alembic.
+revision: str = "6dec9fb5b46c"
+down_revision: Union[str, None] = "61882a919591"
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+
+def upgrade() -> None:
+    # ### commands auto generated by Alembic - please adjust! ###
+    with op.batch_alter_table("meeting", schema=None) as batch_op:
+        batch_op.alter_column("room_id", existing_type=sa.VARCHAR(), nullable=False)
+        batch_op.create_foreign_key(
+            None, "room", ["room_id"], ["id"], ondelete="CASCADE"
+        )
+
+    # ### end Alembic commands ###
+
+
+def downgrade() -> None:
+    # ### commands auto generated by Alembic - please adjust! ###
+    with op.batch_alter_table("meeting", schema=None) as batch_op:
+        batch_op.drop_constraint("meeting_room_id_fkey", type_="foreignkey")
+        batch_op.alter_column("room_id", existing_type=sa.VARCHAR(), nullable=True)
+
+    # ### end Alembic commands ###

server/migrations/versions/6e6ea8e607c5_add_videoplatform_enum_for_rooms_and_.py

@@ -0,0 +1,44 @@
+"""Add VideoPlatform enum for rooms and meetings
+
+Revision ID: 6e6ea8e607c5
+Revises: 61882a919591
+Create Date: 2025-09-02 17:33:21.022214
+
+"""
+
+from typing import Sequence, Union
+
+import sqlalchemy as sa
+from alembic import op
+
+# revision identifiers, used by Alembic.
+revision: str = "6e6ea8e607c5"
+down_revision: Union[str, None] = "61882a919591"
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+
+def upgrade() -> None:
+    # ### commands auto generated by Alembic - please adjust! ###
+    with op.batch_alter_table("meeting", schema=None) as batch_op:
+        batch_op.add_column(
+            sa.Column("platform", sa.String(), server_default="whereby", nullable=False)
+        )
+
+    with op.batch_alter_table("room", schema=None) as batch_op:
+        batch_op.add_column(
+            sa.Column("platform", sa.String(), server_default="whereby", nullable=False)
+        )
+
+    # ### end Alembic commands ###
+
+
+def downgrade() -> None:
+    # ### commands auto generated by Alembic - please adjust! ###
+    with op.batch_alter_table("room", schema=None) as batch_op:
+        batch_op.drop_column("platform")
+
+    with op.batch_alter_table("meeting", schema=None) as batch_op:
+        batch_op.drop_column("platform")
+
+    # ### end Alembic commands ###

server/migrations/versions/def1b5867d4c_make_meeting_room_id_nullable_but_keep_.py

@@ -0,0 +1,34 @@
+"""make meeting room_id nullable but keep foreign key
+
+Revision ID: def1b5867d4c
+Revises: 0ce521cda2ee
+Create Date: 2025-09-11 09:42:18.697264
+
+"""
+
+from typing import Sequence, Union
+
+import sqlalchemy as sa
+from alembic import op
+
+# revision identifiers, used by Alembic.
+revision: str = "def1b5867d4c"
+down_revision: Union[str, None] = "0ce521cda2ee"
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+
+def upgrade() -> None:
+    # ### commands auto generated by Alembic - please adjust! ###
+    with op.batch_alter_table("meeting", schema=None) as batch_op:
+        batch_op.alter_column("room_id", existing_type=sa.VARCHAR(), nullable=True)
+
+    # ### end Alembic commands ###
+
+
+def downgrade() -> None:
+    # ### commands auto generated by Alembic - please adjust! ###
+    with op.batch_alter_table("meeting", schema=None) as batch_op:
+        batch_op.alter_column("room_id", existing_type=sa.VARCHAR(), nullable=False)
+
+    # ### end Alembic commands ###

server/pyproject.toml

@@ -40,6 +40,7 @@ dependencies = [
     "llama-index-llms-openai-like>=0.4.0",
     "pytest-env>=1.1.5",
     "webvtt-py>=0.5.0",
+    "PyJWT>=2.8.0",
 ]
 
 [dependency-groups]

server/reflector/app.py

@@ -12,6 +12,9 @@ from reflector.events import subscribers_shutdown, subscribers_startup
 from reflector.logger import logger
 from reflector.metrics import metrics_init
 from reflector.settings import settings
+from reflector.video_platforms.jitsi import router as jitsi_router
+from reflector.video_platforms.whereby import router as whereby_router
+from reflector.views.jibri_webhook import router as jibri_webhook_router
 from reflector.views.meetings import router as meetings_router
 from reflector.views.rooms import router as rooms_router
 from reflector.views.rtc_offer import router as rtc_offer_router
@@ -26,7 +29,6 @@ from reflector.views.transcripts_upload import router as transcripts_upload_rout
 from reflector.views.transcripts_webrtc import router as transcripts_webrtc_router
 from reflector.views.transcripts_websocket import router as transcripts_websocket_router
 from reflector.views.user import router as user_router
-from reflector.views.whereby import router as whereby_router
 from reflector.views.zulip import router as zulip_router
 
 try:
@@ -86,6 +88,8 @@ app.include_router(transcripts_process_router, prefix="/v1")
 app.include_router(user_router, prefix="/v1")
 app.include_router(zulip_router, prefix="/v1")
 app.include_router(whereby_router, prefix="/v1")
+app.include_router(jitsi_router, prefix="/v1")
+app.include_router(jibri_webhook_router)  # No /v1 prefix, uses /api/v1/jibri
 add_pagination(app)
 
 # prepare celery

server/reflector/asynctask.py

@@ -0,0 +1,27 @@
+import asyncio
+import functools
+
+from reflector.db import get_database
+
+
+def asynctask(f):
+    @functools.wraps(f)
+    def wrapper(*args, **kwargs):
+        async def run_with_db():
+            database = get_database()
+            await database.connect()
+            try:
+                return await f(*args, **kwargs)
+            finally:
+                await database.disconnect()
+
+        coro = run_with_db()
+        try:
+            loop = asyncio.get_running_loop()
+        except RuntimeError:
+            loop = None
+        if loop and loop.is_running():
+            return loop.run_until_complete(coro)
+        return asyncio.run(coro)
+
+    return wrapper

server/reflector/db/meetings.py

@@ -1,12 +1,11 @@
-from datetime import datetime
-from typing import Literal
+from datetime import datetime, timezone
+from typing import Any, Dict, List, Literal
 
 import sqlalchemy as sa
-from fastapi import HTTPException
 from pydantic import BaseModel, Field
 
 from reflector.db import get_database, metadata
-from reflector.db.rooms import Room
+from reflector.db.rooms import Room, VideoPlatform
 from reflector.utils import generate_uuid4
 
 meetings = sa.Table(
@@ -18,8 +17,12 @@ meetings = sa.Table(
     sa.Column("host_room_url", sa.String),
     sa.Column("start_date", sa.DateTime(timezone=True)),
     sa.Column("end_date", sa.DateTime(timezone=True)),
-    sa.Column("user_id", sa.String),
-    sa.Column("room_id", sa.String),
+    sa.Column(
+        "room_id",
+        sa.String,
+        sa.ForeignKey("room.id", ondelete="CASCADE"),
+        nullable=True,
+    ),
     sa.Column("is_locked", sa.Boolean, nullable=False, server_default=sa.false()),
     sa.Column("room_mode", sa.String, nullable=False, server_default="normal"),
     sa.Column("recording_type", sa.String, nullable=False, server_default="cloud"),
@@ -41,6 +44,8 @@ meetings = sa.Table(
         nullable=False,
         server_default=sa.true(),
     ),
+    sa.Column("platform", sa.String, nullable=False, server_default="whereby"),
+    sa.Column("events", sa.JSON, nullable=False, server_default=sa.text("'[]'")),
     sa.Index("idx_meeting_room_id", "room_id"),
     sa.Index(
         "idx_one_active_meeting_per_room",
@@ -54,7 +59,12 @@ meeting_consent = sa.Table(
     "meeting_consent",
     metadata,
     sa.Column("id", sa.String, primary_key=True),
-    sa.Column("meeting_id", sa.String, sa.ForeignKey("meeting.id"), nullable=False),
+    sa.Column(
+        "meeting_id",
+        sa.String,
+        sa.ForeignKey("meeting.id", ondelete="CASCADE"),
+        nullable=False,
+    ),
     sa.Column("user_id", sa.String),
     sa.Column("consent_given", sa.Boolean, nullable=False),
     sa.Column("consent_timestamp", sa.DateTime(timezone=True), nullable=False),
@@ -76,8 +86,7 @@ class Meeting(BaseModel):
     host_room_url: str
     start_date: datetime
     end_date: datetime
-    user_id: str | None = None
-    room_id: str | None = None
+    room_id: str | None
     is_locked: bool = False
     room_mode: Literal["normal", "group"] = "normal"
     recording_type: Literal["none", "local", "cloud"] = "cloud"
@@ -85,6 +94,8 @@ class Meeting(BaseModel):
         "none", "prompt", "automatic", "automatic-2nd-participant"
     ] = "automatic-2nd-participant"
     num_clients: int = 0
+    platform: VideoPlatform = VideoPlatform.WHEREBY
+    events: List[Dict[str, Any]] = Field(default_factory=list)
 
 
 class MeetingController:
@@ -96,12 +107,8 @@ class MeetingController:
         host_room_url: str,
         start_date: datetime,
         end_date: datetime,
-        user_id: str,
         room: Room,
     ):
-        """
-        Create a new meeting
-        """
         meeting = Meeting(
             id=id,
             room_name=room_name,
@@ -109,31 +116,25 @@ class MeetingController:
             host_room_url=host_room_url,
             start_date=start_date,
             end_date=end_date,
-            user_id=user_id,
             room_id=room.id,
             is_locked=room.is_locked,
             room_mode=room.room_mode,
             recording_type=room.recording_type,
             recording_trigger=room.recording_trigger,
+            platform=room.platform,
         )
         query = meetings.insert().values(**meeting.model_dump())
         await get_database().execute(query)
         return meeting
 
     async def get_all_active(self) -> list[Meeting]:
-        """
-        Get active meetings.
-        """
         query = meetings.select().where(meetings.c.is_active)
         return await get_database().fetch_all(query)
 
     async def get_by_room_name(
         self,
         room_name: str,
-    ) -> Meeting:
-        """
-        Get a meeting by room name.
-        """
+    ) -> Meeting | None:
         query = meetings.select().where(meetings.c.room_name == room_name)
         result = await get_database().fetch_one(query)
         if not result:
@@ -141,10 +142,7 @@ class MeetingController:
 
         return Meeting(**result)
 
-    async def get_active(self, room: Room, current_time: datetime) -> Meeting:
-        """
-        Get latest active meeting for a room.
-        """
+    async def get_active(self, room: Room, current_time: datetime) -> Meeting | None:
         end_date = getattr(meetings.c, "end_date")
         query = (
             meetings.select()
@@ -164,36 +162,78 @@ class MeetingController:
         return Meeting(**result)
 
     async def get_by_id(self, meeting_id: str, **kwargs) -> Meeting | None:
-        """
-        Get a meeting by id
-        """
         query = meetings.select().where(meetings.c.id == meeting_id)
         result = await get_database().fetch_one(query)
         if not result:
             return None
         return Meeting(**result)
 
-    async def get_by_id_for_http(self, meeting_id: str, user_id: str | None) -> Meeting:
-        """
-        Get a meeting by ID for HTTP request.
-
-        If not found, it will raise a 404 error.
-        """
-        query = meetings.select().where(meetings.c.id == meeting_id)
-        result = await get_database().fetch_one(query)
-        if not result:
-            raise HTTPException(status_code=404, detail="Meeting not found")
-
-        meeting = Meeting(**result)
-        if result["user_id"] != user_id:
-            meeting.host_room_url = ""
-
-        return meeting
-
     async def update_meeting(self, meeting_id: str, **kwargs):
         query = meetings.update().where(meetings.c.id == meeting_id).values(**kwargs)
         await get_database().execute(query)
 
+    async def add_event(
+        self, meeting_id: str, event_type: str, event_data: Dict[str, Any] = None
+    ):
+        """Add an event to a meeting's events list."""
+        if event_data is None:
+            event_data = {}
+
+        event = {
+            "type": event_type,
+            "timestamp": datetime.now(tz=timezone.utc).isoformat(),
+            "data": event_data,
+        }
+
+        # Get current events
+        query = meetings.select().where(meetings.c.id == meeting_id)
+        result = await get_database().fetch_one(query)
+        if not result:
+            return
+
+        current_events = result["events"] or []
+        current_events.append(event)
+
+        # Update with new events list
+        update_query = (
+            meetings.update()
+            .where(meetings.c.id == meeting_id)
+            .values(events=current_events)
+        )
+        await get_database().execute(update_query)
+
+    async def participant_joined(
+        self, meeting_id: str, participant_data: Dict[str, Any] = None
+    ):
+        """Record a participant joined event."""
+        await self.add_event(meeting_id, "participant_joined", participant_data)
+
+    async def participant_left(
+        self, meeting_id: str, participant_data: Dict[str, Any] = None
+    ):
+        """Record a participant left event."""
+        await self.add_event(meeting_id, "participant_left", participant_data)
+
+    async def recording_started(
+        self, meeting_id: str, recording_data: Dict[str, Any] = None
+    ):
+        """Record a recording started event."""
+        await self.add_event(meeting_id, "recording_started", recording_data)
+
+    async def recording_stopped(
+        self, meeting_id: str, recording_data: Dict[str, Any] = None
+    ):
+        """Record a recording stopped event."""
+        await self.add_event(meeting_id, "recording_stopped", recording_data)
+
+    async def get_events(self, meeting_id: str) -> List[Dict[str, Any]]:
+        """Get all events for a meeting."""
+        query = meetings.select().where(meetings.c.id == meeting_id)
+        result = await get_database().fetch_one(query)
+        if not result:
+            return []
+        return result["events"] or []
+
 
 class MeetingConsentController:
     async def get_by_meeting_id(self, meeting_id: str) -> list[MeetingConsent]:
@@ -214,7 +254,7 @@ class MeetingConsentController:
         result = await get_database().fetch_one(query)
         if result is None:
             return None
-        return MeetingConsent(**result) if result else None
+        return MeetingConsent(**result)
 
     async def upsert(self, consent: MeetingConsent) -> MeetingConsent:
         """Create new consent or update existing one for authenticated users"""

server/reflector/db/rooms.py

@@ -1,4 +1,6 @@
+import secrets
 from datetime import datetime, timezone
+from enum import StrEnum
 from sqlite3 import IntegrityError
 from typing import Literal
 
@@ -10,6 +12,12 @@ from sqlalchemy.sql import false, or_
 from reflector.db import get_database, metadata
 from reflector.utils import generate_uuid4
 
+
+class VideoPlatform(StrEnum):
+    WHEREBY = "whereby"
+    JITSI = "jitsi"
+
+
 rooms = sqlalchemy.Table(
     "room",
     metadata,
@@ -40,6 +48,11 @@ rooms = sqlalchemy.Table(
     sqlalchemy.Column(
         "is_shared", sqlalchemy.Boolean, nullable=False, server_default=false()
     ),
+    sqlalchemy.Column("webhook_url", sqlalchemy.String, nullable=True),
+    sqlalchemy.Column("webhook_secret", sqlalchemy.String, nullable=True),
+    sqlalchemy.Column(
+        "platform", sqlalchemy.String, nullable=False, server_default="whereby"
+    ),
     sqlalchemy.Index("idx_room_is_shared", "is_shared"),
 )
 
@@ -59,6 +72,9 @@ class Room(BaseModel):
         "none", "prompt", "automatic", "automatic-2nd-participant"
     ] = "automatic-2nd-participant"
     is_shared: bool = False
+    webhook_url: str | None = None
+    webhook_secret: str | None = None
+    platform: VideoPlatform = VideoPlatform.WHEREBY
 
 
 class RoomController:
@@ -107,10 +123,16 @@ class RoomController:
         recording_type: str,
         recording_trigger: str,
         is_shared: bool,
+        webhook_url: str = "",
+        webhook_secret: str = "",
+        platform: str = "whereby",
     ):
         """
         Add a new room
         """
+        if webhook_url and not webhook_secret:
+            webhook_secret = secrets.token_urlsafe(32)
+
         room = Room(
             name=name,
             user_id=user_id,
@@ -122,6 +144,9 @@ class RoomController:
             recording_type=recording_type,
             recording_trigger=recording_trigger,
             is_shared=is_shared,
+            webhook_url=webhook_url,
+            webhook_secret=webhook_secret,
+            platform=platform,
         )
         query = rooms.insert().values(**room.model_dump())
         try:
@@ -134,6 +159,9 @@ class RoomController:
         """
         Update a room fields with key/values in values
         """
+        if values.get("webhook_url") and not values.get("webhook_secret"):
+            values["webhook_secret"] = secrets.token_urlsafe(32)
+
         query = rooms.update().where(rooms.c.id == room.id).values(**values)
         try:
             await get_database().execute(query)

server/reflector/db/search.py

@@ -8,12 +8,14 @@ from typing import Annotated, Any, Dict, Iterator
 
 import sqlalchemy
 import webvtt
+from databases.interfaces import Record as DbRecord
 from fastapi import HTTPException
 from pydantic import (
     BaseModel,
     Field,
     NonNegativeFloat,
     NonNegativeInt,
+    TypeAdapter,
     ValidationError,
     constr,
     field_serializer,
@@ -21,9 +23,10 @@ from pydantic import (
 
 from reflector.db import get_database
 from reflector.db.rooms import rooms
-from reflector.db.transcripts import SourceKind, transcripts
+from reflector.db.transcripts import SourceKind, TranscriptStatus, transcripts
 from reflector.db.utils import is_postgresql
 from reflector.logger import logger
+from reflector.utils.string import NonEmptyString, try_parse_non_empty_string
 
 DEFAULT_SEARCH_LIMIT = 20
 SNIPPET_CONTEXT_LENGTH = 50  # Characters before/after match to include
@@ -31,12 +34,13 @@ DEFAULT_SNIPPET_MAX_LENGTH = NonNegativeInt(150)
 DEFAULT_MAX_SNIPPETS = NonNegativeInt(3)
 LONG_SUMMARY_MAX_SNIPPETS = 2
 
-SearchQueryBase = constr(min_length=0, strip_whitespace=True)
+SearchQueryBase = constr(min_length=1, strip_whitespace=True)
 SearchLimitBase = Annotated[int, Field(ge=1, le=100)]
 SearchOffsetBase = Annotated[int, Field(ge=0)]
 SearchTotalBase = Annotated[int, Field(ge=0)]
 
 SearchQuery = Annotated[SearchQueryBase, Field(description="Search query text")]
+search_query_adapter = TypeAdapter(SearchQuery)
 SearchLimit = Annotated[SearchLimitBase, Field(description="Results per page")]
 SearchOffset = Annotated[
     SearchOffsetBase, Field(description="Number of results to skip")
@@ -88,7 +92,7 @@ class WebVTTProcessor:
     @staticmethod
     def generate_snippets(
         webvtt_content: WebVTTContent,
-        query: str,
+        query: SearchQuery,
         max_snippets: NonNegativeInt = DEFAULT_MAX_SNIPPETS,
     ) -> list[str]:
         """Generate snippets from WebVTT content."""
@@ -125,7 +129,7 @@ class SnippetCandidate:
 class SearchParameters(BaseModel):
     """Validated search parameters for full-text search."""
 
-    query_text: SearchQuery
+    query_text: SearchQuery | None = None
     limit: SearchLimit = DEFAULT_SEARCH_LIMIT
     offset: SearchOffset = 0
     user_id: str | None = None
@@ -157,7 +161,7 @@ class SearchResult(BaseModel):
     room_name: str | None = None
     source_kind: SourceKind
     created_at: datetime
-    status: str = Field(..., min_length=1)
+    status: TranscriptStatus = Field(..., min_length=1)
     rank: float = Field(..., ge=0, le=1)
     duration: NonNegativeFloat | None = Field(..., description="Duration in seconds")
     search_snippets: list[str] = Field(
@@ -199,15 +203,13 @@ class SnippetGenerator:
             prev_start = start
 
     @staticmethod
-    def count_matches(text: str, query: str) -> NonNegativeInt:
+    def count_matches(text: str, query: SearchQuery) -> NonNegativeInt:
         """Count total number of matches for a query in text."""
         ZERO = NonNegativeInt(0)
         if not text:
             logger.warning("Empty text for search query in count_matches")
             return ZERO
-        if not query:
-            logger.warning("Empty query for search text in count_matches")
-            return ZERO
+        assert query is not None
         return NonNegativeInt(
             sum(1 for _ in SnippetGenerator.find_all_matches(text, query))
         )
@@ -243,13 +245,14 @@ class SnippetGenerator:
     @staticmethod
     def generate(
         text: str,
-        query: str,
+        query: SearchQuery,
         max_length: NonNegativeInt = DEFAULT_SNIPPET_MAX_LENGTH,
         max_snippets: NonNegativeInt = DEFAULT_MAX_SNIPPETS,
     ) -> list[str]:
         """Generate snippets from text."""
-        if not text or not query:
-            logger.warning("Empty text or query for generate_snippets")
+        assert query is not None
+        if not text:
+            logger.warning("Empty text for generate_snippets")
             return []
 
         candidates = (
@@ -270,7 +273,7 @@ class SnippetGenerator:
     @staticmethod
     def from_summary(
         summary: str,
-        query: str,
+        query: SearchQuery,
         max_snippets: NonNegativeInt = LONG_SUMMARY_MAX_SNIPPETS,
     ) -> list[str]:
         """Generate snippets from summary text."""
@@ -278,9 +281,9 @@ class SnippetGenerator:
 
     @staticmethod
     def combine_sources(
-        summary: str | None,
+        summary: NonEmptyString | None,
         webvtt: WebVTTContent | None,
-        query: str,
+        query: SearchQuery,
         max_total: NonNegativeInt = DEFAULT_MAX_SNIPPETS,
     ) -> tuple[list[str], NonNegativeInt]:
         """Combine snippets from multiple sources and return total match count.
@@ -289,6 +292,11 @@ class SnippetGenerator:
 
         snippets can be empty for real in case of e.g. title match
         """
+
+        assert (
+            summary is not None or webvtt is not None
+        ), "At least one source must be present"
+
         webvtt_matches = 0
         summary_matches = 0
 
@@ -355,8 +363,8 @@ class SearchController:
                 else_=rooms.c.name,
             ).label("room_name"),
         ]
-
-        if params.query_text:
+        search_query = None
+        if params.query_text is not None:
             search_query = sqlalchemy.func.websearch_to_tsquery(
                 "english", params.query_text
             )
@@ -373,7 +381,9 @@ class SearchController:
             transcripts.join(rooms, transcripts.c.room_id == rooms.c.id, isouter=True)
         )
 
-        if params.query_text:
+        if params.query_text is not None:
+            # because already initialized based on params.query_text presence above
+            assert search_query is not None
             base_query = base_query.where(
                 transcripts.c.search_vector_en.op("@@")(search_query)
             )
@@ -393,7 +403,7 @@ class SearchController:
                 transcripts.c.source_kind == params.source_kind
             )
 
-        if params.query_text:
+        if params.query_text is not None:
             order_by = sqlalchemy.desc(sqlalchemy.text("rank"))
         else:
             order_by = sqlalchemy.desc(transcripts.c.created_at)
@@ -407,19 +417,29 @@ class SearchController:
         )
         total = await get_database().fetch_val(count_query)
 
-        def _process_result(r) -> SearchResult:
+        def _process_result(r: DbRecord) -> SearchResult:
             r_dict: Dict[str, Any] = dict(r)
+
             webvtt_raw: str | None = r_dict.pop("webvtt", None)
+            webvtt: WebVTTContent | None
             if webvtt_raw:
                 webvtt = WebVTTProcessor.parse(webvtt_raw)
             else:
                 webvtt = None
-            long_summary: str | None = r_dict.pop("long_summary", None)
+
+            long_summary_r: str | None = r_dict.pop("long_summary", None)
+            long_summary: NonEmptyString = try_parse_non_empty_string(long_summary_r)
             room_name: str | None = r_dict.pop("room_name", None)
             db_result = SearchResultDB.model_validate(r_dict)
 
-            snippets, total_match_count = SnippetGenerator.combine_sources(
-                long_summary, webvtt, params.query_text, DEFAULT_MAX_SNIPPETS
+            at_least_one_source = webvtt is not None or long_summary is not None
+            has_query = params.query_text is not None
+            snippets, total_match_count = (
+                SnippetGenerator.combine_sources(
+                    long_summary, webvtt, params.query_text, DEFAULT_MAX_SNIPPETS
+                )
+                if has_query and at_least_one_source
+                else ([], 0)
             )
 
             return SearchResult(

server/reflector/db/transcripts.py

@@ -122,6 +122,15 @@ def generate_transcript_name() -> str:
     return f"Transcript {now.strftime('%Y-%m-%d %H:%M:%S')}"
 
 
+TranscriptStatus = Literal[
+    "idle", "uploaded", "recording", "processing", "error", "ended"
+]
+
+
+class StrValue(BaseModel):
+    value: str
+
+
 class AudioWaveform(BaseModel):
     data: list[float]
 
@@ -185,7 +194,7 @@ class Transcript(BaseModel):
     id: str = Field(default_factory=generate_uuid4)
     user_id: str | None = None
     name: str = Field(default_factory=generate_transcript_name)
-    status: str = "idle"
+    status: TranscriptStatus = "idle"
     duration: float = 0
     created_at: datetime = Field(default_factory=lambda: datetime.now(timezone.utc))
     title: str | None = None
@@ -732,5 +741,27 @@ class TranscriptController:
         transcript.delete_participant(participant_id)
         await self.update(transcript, {"participants": transcript.participants_dump()})
 
+    async def set_status(
+        self, transcript_id: str, status: TranscriptStatus
+    ) -> TranscriptEvent | None:
+        """
+        Update the status of a transcript
+
+        Will add an event STATUS + update the status field of transcript
+        """
+        async with self.transaction():
+            transcript = await self.get_by_id(transcript_id)
+            if not transcript:
+                raise Exception(f"Transcript {transcript_id} not found")
+            if transcript.status == status:
+                return
+            resp = await self.append_event(
+                transcript=transcript,
+                event="STATUS",
+                data=StrValue(value=status),
+            )
+            await self.update(transcript, {"status": status})
+        return resp
+
 
 transcripts_controller = TranscriptController()

server/reflector/jibri_events.py

@@ -0,0 +1,227 @@
+import json
+from pathlib import Path
+from typing import Any, Dict, List, Literal, Optional, Union
+
+from pydantic import BaseModel
+from typing_extensions import TypedDict
+
+
+class ParticipantInfo(BaseModel):
+    jid: str
+    nick: str
+    id: str
+    is_moderator: bool = False
+
+
+class ParticipantLeftInfo(BaseModel):
+    jid: str
+    nick: Optional[str] = None
+    duration_seconds: Optional[int] = None
+
+
+class RoomCreatedEvent(BaseModel):
+    type: Literal["room_created"]
+    timestamp: int
+    room_name: str
+    room_jid: str
+    meeting_url: str
+
+
+class RecordingStartedEvent(BaseModel):
+    type: Literal["recording_started"]
+    timestamp: int
+    room_name: str
+    session_id: str
+    jibri_jid: str
+
+
+class RecordingStoppedEvent(BaseModel):
+    type: Literal["recording_stopped"]
+    timestamp: int
+    room_name: str
+    session_id: str
+    meeting_url: str
+
+
+class ParticipantJoinedEvent(BaseModel):
+    type: Literal["participant_joined"]
+    timestamp: int
+    room_name: str
+    participant: ParticipantInfo
+
+
+class ParticipantLeftEvent(BaseModel):
+    type: Literal["participant_left"]
+    timestamp: int
+    room_name: str
+    participant: ParticipantLeftInfo
+
+
+class SpeakerActiveEvent(BaseModel):
+    type: Literal["speaker_active"]
+    timestamp: int
+    room_name: str
+    speaker_jid: str
+    speaker_nick: str
+    duration: int
+
+
+class DominantSpeakerChangedEvent(BaseModel):
+    type: Literal["dominant_speaker_changed"]
+    timestamp: int
+    room_name: str
+    previous: str
+    current: str
+
+
+JitsiEvent = Union[
+    RoomCreatedEvent,
+    RecordingStartedEvent,
+    RecordingStoppedEvent,
+    ParticipantJoinedEvent,
+    ParticipantLeftEvent,
+    SpeakerActiveEvent,
+    DominantSpeakerChangedEvent,
+]
+
+
+class RoomInfo(TypedDict):
+    name: str
+    jid: str
+    created_at: int
+    meeting_url: str
+    recording_stopped_at: Optional[int]
+
+
+class ParticipantData(TypedDict):
+    jid: str
+    nick: str
+    id: str
+    is_moderator: bool
+    joined_at: int
+    left_at: Optional[int]
+    duration: Optional[int]
+    events: List[str]
+
+
+class SpeakerStats(TypedDict):
+    total_time: int
+    nick: str
+
+
+class ParsedMetadata(TypedDict):
+    room: RoomInfo
+    participants: List[ParticipantData]
+    speaker_stats: Dict[str, SpeakerStats]
+    event_count: int
+
+
+class JitsiEventParser:
+    def parse_event(self, event_data: Dict[str, Any]) -> Optional[JitsiEvent]:
+        event_type = event_data.get("type")
+
+        try:
+            if event_type == "room_created":
+                return RoomCreatedEvent(**event_data)
+            elif event_type == "recording_started":
+                return RecordingStartedEvent(**event_data)
+            elif event_type == "recording_stopped":
+                return RecordingStoppedEvent(**event_data)
+            elif event_type == "participant_joined":
+                return ParticipantJoinedEvent(**event_data)
+            elif event_type == "participant_left":
+                return ParticipantLeftEvent(**event_data)
+            elif event_type == "speaker_active":
+                return SpeakerActiveEvent(**event_data)
+            elif event_type == "dominant_speaker_changed":
+                return DominantSpeakerChangedEvent(**event_data)
+            else:
+                return None
+        except Exception:
+            return None
+
+    def parse_events_file(self, recording_path: str) -> ParsedMetadata:
+        events_file = Path(recording_path) / "events.jsonl"
+
+        room_info: RoomInfo = {
+            "name": "",
+            "jid": "",
+            "created_at": 0,
+            "meeting_url": "",
+            "recording_stopped_at": None,
+        }
+
+        if not events_file.exists():
+            return ParsedMetadata(
+                room=room_info, participants=[], speaker_stats={}, event_count=0
+            )
+
+        events: List[JitsiEvent] = []
+        participants: Dict[str, ParticipantData] = {}
+        speaker_stats: Dict[str, SpeakerStats] = {}
+
+        with open(events_file, "r") as f:
+            for line in f:
+                if not line.strip():
+                    continue
+
+                try:
+                    event_data = json.loads(line)
+                    event = self.parse_event(event_data)
+
+                    if event is None:
+                        continue
+
+                    events.append(event)
+
+                    if isinstance(event, RoomCreatedEvent):
+                        room_info = {
+                            "name": event.room_name,
+                            "jid": event.room_jid,
+                            "created_at": event.timestamp,
+                            "meeting_url": event.meeting_url,
+                            "recording_stopped_at": None,
+                        }
+
+                    elif isinstance(event, ParticipantJoinedEvent):
+                        participants[event.participant.id] = {
+                            "jid": event.participant.jid,
+                            "nick": event.participant.nick,
+                            "id": event.participant.id,
+                            "is_moderator": event.participant.is_moderator,
+                            "joined_at": event.timestamp,
+                            "left_at": None,
+                            "duration": None,
+                            "events": ["joined"],
+                        }
+
+                    elif isinstance(event, ParticipantLeftEvent):
+                        participant_id = event.participant.jid.split("/")[0]
+                        if participant_id in participants:
+                            participants[participant_id]["left_at"] = event.timestamp
+                            participants[participant_id]["duration"] = (
+                                event.participant.duration_seconds
+                            )
+                            participants[participant_id]["events"].append("left")
+
+                    elif isinstance(event, SpeakerActiveEvent):
+                        if event.speaker_jid not in speaker_stats:
+                            speaker_stats[event.speaker_jid] = {
+                                "total_time": 0,
+                                "nick": event.speaker_nick,
+                            }
+                        speaker_stats[event.speaker_jid]["total_time"] += event.duration
+
+                    elif isinstance(event, RecordingStoppedEvent):
+                        room_info["recording_stopped_at"] = event.timestamp
+                        room_info["meeting_url"] = event.meeting_url
+
+                except (json.JSONDecodeError, Exception):
+                    continue
+
+        return ParsedMetadata(
+            room=room_info,
+            participants=list(participants.values()),
+            speaker_stats=speaker_stats,
+            event_count=len(events),
+        )

server/reflector/pipelines/main_file_pipeline.py

@@ -7,18 +7,28 @@ Uses parallel processing for transcription, diarization, and waveform generation
 """
 
 import asyncio
+import uuid
 from pathlib import Path
 
 import av
 import structlog
-from celery import shared_task
+from celery import chain, shared_task
 
+from reflector.asynctask import asynctask
+from reflector.db.rooms import rooms_controller
 from reflector.db.transcripts import (
+    SourceKind,
     Transcript,
+    TranscriptStatus,
     transcripts_controller,
 )
 from reflector.logger import logger
-from reflector.pipelines.main_live_pipeline import PipelineMainBase, asynctask
+from reflector.pipelines.main_live_pipeline import (
+    PipelineMainBase,
+    broadcast_to_sockets,
+    task_cleanup_consent,
+    task_pipeline_post_to_zulip,
+)
 from reflector.processors import (
     AudioFileWriterProcessor,
     TranscriptFinalSummaryProcessor,
@@ -43,6 +53,7 @@ from reflector.processors.types import (
 )
 from reflector.settings import settings
 from reflector.storage import get_transcripts_storage
+from reflector.worker.webhook import send_transcript_webhook
 
 
 class EmptyPipeline:
@@ -83,12 +94,27 @@ class PipelineMainFile(PipelineMainBase):
                 exc_info=result,
             )
 
+    @broadcast_to_sockets
+    async def set_status(self, transcript_id: str, status: TranscriptStatus):
+        async with self.lock_transaction():
+            return await transcripts_controller.set_status(transcript_id, status)
+
     async def process(self, file_path: Path):
         """Main entry point for file processing"""
         self.logger.info(f"Starting file pipeline for {file_path}")
 
         transcript = await self.get_transcript()
 
+        # Clear transcript as we're going to regenerate everything
+        async with self.transaction():
+            await transcripts_controller.update(
+                transcript,
+                {
+                    "events": [],
+                    "topics": [],
+                },
+            )
+
         # Extract audio and write to transcript location
         audio_path = await self.extract_and_write_audio(file_path, transcript)
 
@@ -105,6 +131,8 @@ class PipelineMainFile(PipelineMainBase):
 
         self.logger.info("File pipeline complete")
 
+        await transcripts_controller.set_status(transcript.id, "ended")
+
     async def extract_and_write_audio(
         self, file_path: Path, transcript: Transcript
     ) -> Path:
@@ -353,6 +381,28 @@ class PipelineMainFile(PipelineMainBase):
         await processor.flush()
 
 
+@shared_task
+@asynctask
+async def task_send_webhook_if_needed(*, transcript_id: str):
+    """Send webhook if this is a room recording with webhook configured"""
+    transcript = await transcripts_controller.get_by_id(transcript_id)
+    if not transcript:
+        return
+
+    if transcript.source_kind == SourceKind.ROOM and transcript.room_id:
+        room = await rooms_controller.get_by_id(transcript.room_id)
+        if room and room.webhook_url:
+            logger.info(
+                "Dispatching webhook",
+                transcript_id=transcript_id,
+                room_id=room.id,
+                webhook_url=room.webhook_url,
+            )
+            send_transcript_webhook.delay(
+                transcript_id, room.id, event_id=uuid.uuid4().hex
+            )
+
+
 @shared_task
 @asynctask
 async def task_pipeline_file_process(*, transcript_id: str):
@@ -362,14 +412,28 @@ async def task_pipeline_file_process(*, transcript_id: str):
     if not transcript:
         raise Exception(f"Transcript {transcript_id} not found")
 
-    # Find the file to process
-    audio_file = next(transcript.data_path.glob("upload.*"), None)
-    if not audio_file:
-        audio_file = next(transcript.data_path.glob("audio.*"), None)
-
-    if not audio_file:
-        raise Exception("No audio file found to process")
-
-    # Run file pipeline
     pipeline = PipelineMainFile(transcript_id=transcript_id)
-    await pipeline.process(audio_file)
+    try:
+        await pipeline.set_status(transcript_id, "processing")
+
+        # Find the file to process
+        audio_file = next(transcript.data_path.glob("upload.*"), None)
+        if not audio_file:
+            audio_file = next(transcript.data_path.glob("audio.*"), None)
+
+        if not audio_file:
+            raise Exception("No audio file found to process")
+
+        await pipeline.process(audio_file)
+
+    except Exception:
+        await pipeline.set_status(transcript_id, "error")
+        raise
+
+    # Run post-processing chain: consent cleanup -> zulip -> webhook
+    post_chain = chain(
+        task_cleanup_consent.si(transcript_id=transcript_id),
+        task_pipeline_post_to_zulip.si(transcript_id=transcript_id),
+        task_send_webhook_if_needed.si(transcript_id=transcript_id),
+    )
+    post_chain.delay()

server/reflector/pipelines/main_live_pipeline.py

@@ -22,7 +22,7 @@ from celery import chord, current_task, group, shared_task
 from pydantic import BaseModel
 from structlog import BoundLogger as Logger
 
-from reflector.db import get_database
+from reflector.asynctask import asynctask
 from reflector.db.meetings import meeting_consent_controller, meetings_controller
 from reflector.db.recordings import recordings_controller
 from reflector.db.rooms import rooms_controller
@@ -32,6 +32,7 @@ from reflector.db.transcripts import (
     TranscriptFinalLongSummary,
     TranscriptFinalShortSummary,
     TranscriptFinalTitle,
+    TranscriptStatus,
     TranscriptText,
     TranscriptTopic,
     TranscriptWaveform,
@@ -40,8 +41,9 @@ from reflector.db.transcripts import (
 from reflector.logger import logger
 from reflector.pipelines.runner import PipelineMessage, PipelineRunner
 from reflector.processors import (
-    AudioChunkerProcessor,
+    AudioChunkerAutoProcessor,
     AudioDiarizationAutoProcessor,
+    AudioDownscaleProcessor,
     AudioFileWriterProcessor,
     AudioMergeProcessor,
     AudioTranscriptAutoProcessor,
@@ -68,29 +70,6 @@ from reflector.zulip import (
 )
 
 
-def asynctask(f):
-    @functools.wraps(f)
-    def wrapper(*args, **kwargs):
-        async def run_with_db():
-            database = get_database()
-            await database.connect()
-            try:
-                return await f(*args, **kwargs)
-            finally:
-                await database.disconnect()
-
-        coro = run_with_db()
-        try:
-            loop = asyncio.get_running_loop()
-        except RuntimeError:
-            loop = None
-        if loop and loop.is_running():
-            return loop.run_until_complete(coro)
-        return asyncio.run(coro)
-
-    return wrapper
-
-
 def broadcast_to_sockets(func):
     """
     Decorator to broadcast transcript event to websockets
@@ -187,8 +166,15 @@ class PipelineMainBase(PipelineRunner[PipelineMessage], Generic[PipelineMessage]
         ]
 
     @asynccontextmanager
-    async def transaction(self):
+    async def lock_transaction(self):
+        # This lock is to prevent multiple processor starting adding
+        # into event array at the same time
         async with self._lock:
+            yield
+
+    @asynccontextmanager
+    async def transaction(self):
+        async with self.lock_transaction():
             async with transcripts_controller.transaction():
                 yield
 
@@ -197,14 +183,14 @@ class PipelineMainBase(PipelineRunner[PipelineMessage], Generic[PipelineMessage]
         # if it's the first part, update the status of the transcript
         # but do not set the ended status yet.
         if isinstance(self, PipelineMainLive):
-            status_mapping = {
+            status_mapping: dict[str, TranscriptStatus] = {
                 "started": "recording",
                 "push": "recording",
                 "flush": "processing",
                 "error": "error",
             }
         elif isinstance(self, PipelineMainFinalSummaries):
-            status_mapping = {
+            status_mapping: dict[str, TranscriptStatus] = {
                 "push": "processing",
                 "flush": "processing",
                 "error": "error",
@@ -220,22 +206,8 @@ class PipelineMainBase(PipelineRunner[PipelineMessage], Generic[PipelineMessage]
             return
 
         # when the status of the pipeline changes, update the transcript
-        async with self.transaction():
-            transcript = await self.get_transcript()
-            if status == transcript.status:
-                return
-            resp = await transcripts_controller.append_event(
-                transcript=transcript,
-                event="STATUS",
-                data=StrValue(value=status),
-            )
-            await transcripts_controller.update(
-                transcript,
-                {
-                    "status": status,
-                },
-            )
-            return resp
+        async with self._lock:
+            return await transcripts_controller.set_status(self.transcript_id, status)
 
     @broadcast_to_sockets
     async def on_transcript(self, data):
@@ -365,7 +337,8 @@ class PipelineMainLive(PipelineMainBase):
                 path=transcript.audio_wav_filename,
                 on_duration=self.on_duration,
             ),
-            AudioChunkerProcessor(),
+            AudioDownscaleProcessor(),
+            AudioChunkerAutoProcessor(),
             AudioMergeProcessor(),
             AudioTranscriptAutoProcessor.as_threaded(),
             TranscriptLinerProcessor(),
@@ -792,7 +765,7 @@ def pipeline_post(*, transcript_id: str):
         chain_final_summaries,
     ) | task_pipeline_post_to_zulip.si(transcript_id=transcript_id)
 
-    chain.delay()
+    return chain.delay()
 
 
 @get_transcript

server/reflector/processors/__init__.py

@@ -1,5 +1,7 @@
 from .audio_chunker import AudioChunkerProcessor  # noqa: F401
+from .audio_chunker_auto import AudioChunkerAutoProcessor  # noqa: F401
 from .audio_diarization_auto import AudioDiarizationAutoProcessor  # noqa: F401
+from .audio_downscale import AudioDownscaleProcessor  # noqa: F401
 from .audio_file_writer import AudioFileWriterProcessor  # noqa: F401
 from .audio_merge import AudioMergeProcessor  # noqa: F401
 from .audio_transcript import AudioTranscriptProcessor  # noqa: F401

server/reflector/processors/audio_chunker.py

@@ -1,340 +1,78 @@
 from typing import Optional
 
 import av
-import numpy as np
-import torch
-from silero_vad import VADIterator, load_silero_vad
+from prometheus_client import Counter, Histogram
 
 from reflector.processors.base import Processor
 
 
 class AudioChunkerProcessor(Processor):
     """
-    Assemble audio frames into chunks with VAD-based speech detection
+    Base class for assembling audio frames into chunks
     """
 
     INPUT_TYPE = av.AudioFrame
     OUTPUT_TYPE = list[av.AudioFrame]
 
-    def __init__(
-        self,
-        block_frames=256,
-        max_frames=1024,
-        vad_threshold=0.5,
-        use_onnx=False,
-        min_frames=2,
-    ):
-        super().__init__()
+    m_chunk = Histogram(
+        "audio_chunker",
+        "Time spent in AudioChunker.chunk",
+        ["backend"],
+    )
+    m_chunk_call = Counter(
+        "audio_chunker_call",
+        "Number of calls to AudioChunker.chunk",
+        ["backend"],
+    )
+    m_chunk_success = Counter(
+        "audio_chunker_success",
+        "Number of successful calls to AudioChunker.chunk",
+        ["backend"],
+    )
+    m_chunk_failure = Counter(
+        "audio_chunker_failure",
+        "Number of failed calls to AudioChunker.chunk",
+        ["backend"],
+    )
+
+    def __init__(self, *args, **kwargs):
+        name = self.__class__.__name__
+        self.m_chunk = self.m_chunk.labels(name)
+        self.m_chunk_call = self.m_chunk_call.labels(name)
+        self.m_chunk_success = self.m_chunk_success.labels(name)
+        self.m_chunk_failure = self.m_chunk_failure.labels(name)
+        super().__init__(*args, **kwargs)
         self.frames: list[av.AudioFrame] = []
-        self.block_frames = block_frames
-        self.max_frames = max_frames
-        self.vad_threshold = vad_threshold
-        self.min_frames = min_frames
-
-        # Initialize Silero VAD
-        self._init_vad(use_onnx)
-
-    def _init_vad(self, use_onnx=False):
-        """Initialize Silero VAD model"""
-        try:
-            torch.set_num_threads(1)
-            self.vad_model = load_silero_vad(onnx=use_onnx)
-            self.vad_iterator = VADIterator(self.vad_model, sampling_rate=16000)
-            self.logger.info("Silero VAD initialized successfully")
-
-        except Exception as e:
-            self.logger.error(f"Failed to initialize Silero VAD: {e}")
-            self.vad_model = None
-            self.vad_iterator = None
 
     async def _push(self, data: av.AudioFrame):
-        self.frames.append(data)
-        # print("timestamp", data.pts * data.time_base * 1000)
-
-        # Check for speech segments every 32 frames (~1 second)
-        if len(self.frames) >= 32 and len(self.frames) % 32 == 0:
-            await self._process_block()
-
-        # Safety fallback - emit if we hit max frames
-        elif len(self.frames) >= self.max_frames:
-            self.logger.warning(
-                f"AudioChunkerProcessor: Reached max frames ({self.max_frames}), "
-                f"emitting first {self.max_frames // 2} frames"
-            )
-            frames_to_emit = self.frames[: self.max_frames // 2]
-            self.frames = self.frames[self.max_frames // 2 :]
-            if len(frames_to_emit) >= self.min_frames:
-                await self.emit(frames_to_emit)
-            else:
-                self.logger.debug(
-                    f"Ignoring fallback segment with {len(frames_to_emit)} frames "
-                    f"(< {self.min_frames} minimum)"
+        """Process incoming audio frame"""
+        # Validate audio format on first frame
+        if len(self.frames) == 0:
+            if data.sample_rate != 16000 or len(data.layout.channels) != 1:
+                raise ValueError(
+                    f"AudioChunkerProcessor expects 16kHz mono audio, got {data.sample_rate}Hz "
+                    f"with {len(data.layout.channels)} channel(s). "
+                    f"Use AudioDownscaleProcessor before this processor."
                 )
 
-    async def _process_block(self):
-        # Need at least 32 frames for VAD detection (~1 second)
-        if len(self.frames) < 32 or self.vad_iterator is None:
-            return
-
-        # Processing block with current buffer size
-        # print(f"Processing block: {len(self.frames)} frames in buffer")
-
         try:
-            # Convert frames to numpy array for VAD
-            audio_array = self._frames_to_numpy(self.frames)
+            self.m_chunk_call.inc()
+            with self.m_chunk.time():
+                result = await self._chunk(data)
+            self.m_chunk_success.inc()
+            if result:
+                await self.emit(result)
+        except Exception:
+            self.m_chunk_failure.inc()
+            raise
 
-            if audio_array is None:
-                # Fallback: emit all frames if conversion failed
-                frames_to_emit = self.frames[:]
-                self.frames = []
-                if len(frames_to_emit) >= self.min_frames:
-                    await self.emit(frames_to_emit)
-                else:
-                    self.logger.debug(
-                        f"Ignoring conversion-failed segment with {len(frames_to_emit)} frames "
-                        f"(< {self.min_frames} minimum)"
-                    )
-                return
-
-            # Find complete speech segments in the buffer
-            speech_end_frame = self._find_speech_segment_end(audio_array)
-
-            if speech_end_frame is None or speech_end_frame <= 0:
-                # No speech found but buffer is getting large
-                if len(self.frames) > 512:
-                    # Check if it's all silence and can be discarded
-                    # No speech segment found, buffer at {len(self.frames)} frames
-
-                    # Could emit silence or discard old frames here
-                    # For now, keep first 256 frames and discard older silence
-                    if len(self.frames) > 768:
-                        self.logger.debug(
-                            f"Discarding {len(self.frames) - 256} old frames (likely silence)"
-                        )
-                        self.frames = self.frames[-256:]
-                return
-
-            # Calculate segment timing information
-            frames_to_emit = self.frames[:speech_end_frame]
-
-            # Get timing from av.AudioFrame
-            if frames_to_emit:
-                first_frame = frames_to_emit[0]
-                last_frame = frames_to_emit[-1]
-                sample_rate = first_frame.sample_rate
-
-                # Calculate duration
-                total_samples = sum(f.samples for f in frames_to_emit)
-                duration_seconds = total_samples / sample_rate if sample_rate > 0 else 0
-
-                # Get timestamps if available
-                start_time = (
-                    first_frame.pts * first_frame.time_base if first_frame.pts else 0
-                )
-                end_time = (
-                    last_frame.pts * last_frame.time_base if last_frame.pts else 0
-                )
-
-                # Convert to HH:MM:SS format for logging
-                def format_time(seconds):
-                    if not seconds:
-                        return "00:00:00"
-                    total_seconds = int(float(seconds))
-                    hours = total_seconds // 3600
-                    minutes = (total_seconds % 3600) // 60
-                    secs = total_seconds % 60
-                    return f"{hours:02d}:{minutes:02d}:{secs:02d}"
-
-                start_formatted = format_time(start_time)
-                end_formatted = format_time(end_time)
-
-                # Keep remaining frames for next processing
-                remaining_after = len(self.frames) - speech_end_frame
-
-                # Single structured log line
-                self.logger.info(
-                    "Speech segment found",
-                    start=start_formatted,
-                    end=end_formatted,
-                    frames=speech_end_frame,
-                    duration=round(duration_seconds, 2),
-                    buffer_before=len(self.frames),
-                    remaining=remaining_after,
-                )
-
-            # Keep remaining frames for next processing
-            self.frames = self.frames[speech_end_frame:]
-
-            # Filter out segments with too few frames
-            if len(frames_to_emit) >= self.min_frames:
-                await self.emit(frames_to_emit)
-            else:
-                self.logger.debug(
-                    f"Ignoring segment with {len(frames_to_emit)} frames "
-                    f"(< {self.min_frames} minimum)"
-                )
-
-        except Exception as e:
-            self.logger.error(f"Error in VAD processing: {e}")
-            # Fallback to simple chunking
-            if len(self.frames) >= self.block_frames:
-                frames_to_emit = self.frames[: self.block_frames]
-                self.frames = self.frames[self.block_frames :]
-                if len(frames_to_emit) >= self.min_frames:
-                    await self.emit(frames_to_emit)
-                else:
-                    self.logger.debug(
-                        f"Ignoring exception-fallback segment with {len(frames_to_emit)} frames "
-                        f"(< {self.min_frames} minimum)"
-                    )
-
-    def _frames_to_numpy(self, frames: list[av.AudioFrame]) -> Optional[np.ndarray]:
-        """Convert av.AudioFrame list to numpy array for VAD processing"""
-        if not frames:
-            return None
-
-        try:
-            first_frame = frames[0]
-            original_sample_rate = first_frame.sample_rate
-
-            audio_data = []
-            for frame in frames:
-                frame_array = frame.to_ndarray()
-
-                # Handle stereo -> mono conversion
-                if len(frame_array.shape) == 2 and frame_array.shape[0] > 1:
-                    frame_array = np.mean(frame_array, axis=0)
-                elif len(frame_array.shape) == 2:
-                    frame_array = frame_array.flatten()
-
-                audio_data.append(frame_array)
-
-            if not audio_data:
-                return None
-
-            combined_audio = np.concatenate(audio_data)
-
-            # Resample from 48kHz to 16kHz if needed
-            if original_sample_rate != 16000:
-                combined_audio = self._resample_audio(
-                    combined_audio, original_sample_rate, 16000
-                )
-
-            # Ensure float32 format
-            if combined_audio.dtype == np.int16:
-                # Normalize int16 audio to float32 in range [-1.0, 1.0]
-                combined_audio = combined_audio.astype(np.float32) / 32768.0
-            elif combined_audio.dtype != np.float32:
-                combined_audio = combined_audio.astype(np.float32)
-
-            return combined_audio
-
-        except Exception as e:
-            self.logger.error(f"Error converting frames to numpy: {e}")
-
-        return None
-
-    def _resample_audio(
-        self, audio: np.ndarray, from_sr: int, to_sr: int
-    ) -> np.ndarray:
-        """Simple linear resampling from from_sr to to_sr"""
-        if from_sr == to_sr:
-            return audio
-
-        try:
-            # Simple linear interpolation resampling
-            ratio = to_sr / from_sr
-            new_length = int(len(audio) * ratio)
-
-            # Create indices for interpolation
-            old_indices = np.linspace(0, len(audio) - 1, new_length)
-            resampled = np.interp(old_indices, np.arange(len(audio)), audio)
-
-            return resampled.astype(np.float32)
-
-        except Exception as e:
-            self.logger.error("Resampling error", exc_info=e)
-            # Fallback: simple decimation/repetition
-            if from_sr > to_sr:
-                # Downsample by taking every nth sample
-                step = from_sr // to_sr
-                return audio[::step]
-            else:
-                # Upsample by repeating samples
-                repeat = to_sr // from_sr
-                return np.repeat(audio, repeat)
-
-    def _find_speech_segment_end(self, audio_array: np.ndarray) -> Optional[int]:
-        """Find complete speech segments and return frame index at segment end"""
-        if self.vad_iterator is None or len(audio_array) == 0:
-            return None
-
-        try:
-            # Process audio in 512-sample windows for VAD
-            window_size = 512
-            min_silence_windows = 3  # Require 3 windows of silence after speech
-
-            # Track speech state
-            in_speech = False
-            speech_start = None
-            speech_end = None
-            silence_count = 0
-
-            for i in range(0, len(audio_array), window_size):
-                chunk = audio_array[i : i + window_size]
-                if len(chunk) < window_size:
-                    chunk = np.pad(chunk, (0, window_size - len(chunk)))
-
-                # Detect if this window has speech
-                speech_dict = self.vad_iterator(chunk, return_seconds=True)
-
-                # VADIterator returns dict with 'start' and 'end' when speech segments are detected
-                if speech_dict:
-                    if not in_speech:
-                        # Speech started
-                        speech_start = i
-                        in_speech = True
-                        # Debug: print(f"Speech START at sample {i}, VAD: {speech_dict}")
-                    silence_count = 0  # Reset silence counter
-                    continue
-
-                if not in_speech:
-                    continue
-
-                # We're in speech but found silence
-                silence_count += 1
-                if silence_count < min_silence_windows:
-                    continue
-
-                # Found end of speech segment
-                speech_end = i - (min_silence_windows - 1) * window_size
-                # Debug: print(f"Speech END at sample {speech_end}")
-
-                # Convert sample position to frame index
-                samples_per_frame = self.frames[0].samples if self.frames else 1024
-                # Account for resampling: we process at 16kHz but frames might be 48kHz
-                resample_ratio = 48000 / 16000  # 3x
-                actual_sample_pos = int(speech_end * resample_ratio)
-                frame_index = actual_sample_pos // samples_per_frame
-
-                # Ensure we don't exceed buffer
-                frame_index = min(frame_index, len(self.frames))
-                return frame_index
-
-            return None
-
-        except Exception as e:
-            self.logger.error(f"Error finding speech segment: {e}")
-            return None
+    async def _chunk(self, data: av.AudioFrame) -> Optional[list[av.AudioFrame]]:
+        """
+        Process audio frame and return chunk when ready.
+        Subclasses should implement their chunking logic here.
+        """
+        raise NotImplementedError
 
     async def _flush(self):
-        frames = self.frames[:]
-        self.frames = []
-        if frames:
-            if len(frames) >= self.min_frames:
-                await self.emit(frames)
-            else:
-                self.logger.debug(
-                    f"Ignoring flush segment with {len(frames)} frames "
-                    f"(< {self.min_frames} minimum)"
-                )
+        """Flush any remaining frames when processing ends"""
+        raise NotImplementedError

server/reflector/processors/audio_chunker_auto.py

@@ -0,0 +1,32 @@
+import importlib
+
+from reflector.processors.audio_chunker import AudioChunkerProcessor
+from reflector.settings import settings
+
+
+class AudioChunkerAutoProcessor(AudioChunkerProcessor):
+    _registry = {}
+
+    @classmethod
+    def register(cls, name, kclass):
+        cls._registry[name] = kclass
+
+    def __new__(cls, name: str | None = None, **kwargs):
+        if name is None:
+            name = settings.AUDIO_CHUNKER_BACKEND
+        if name not in cls._registry:
+            module_name = f"reflector.processors.audio_chunker_{name}"
+            importlib.import_module(module_name)
+
+        # gather specific configuration for the processor
+        # search `AUDIO_CHUNKER_BACKEND_XXX_YYY`, push to constructor as `backend_xxx_yyy`
+        config = {}
+        name_upper = name.upper()
+        settings_prefix = "AUDIO_CHUNKER_"
+        config_prefix = f"{settings_prefix}{name_upper}_"
+        for key, value in settings:
+            if key.startswith(config_prefix):
+                config_name = key[len(settings_prefix) :].lower()
+                config[config_name] = value
+
+        return cls._registry[name](**config | kwargs)

server/reflector/processors/audio_chunker_frames.py

@@ -0,0 +1,34 @@
+from typing import Optional
+
+import av
+
+from reflector.processors.audio_chunker import AudioChunkerProcessor
+from reflector.processors.audio_chunker_auto import AudioChunkerAutoProcessor
+
+
+class AudioChunkerFramesProcessor(AudioChunkerProcessor):
+    """
+    Simple frame-based audio chunker that emits chunks after a fixed number of frames
+    """
+
+    def __init__(self, max_frames=256, **kwargs):
+        super().__init__(**kwargs)
+        self.max_frames = max_frames
+
+    async def _chunk(self, data: av.AudioFrame) -> Optional[list[av.AudioFrame]]:
+        self.frames.append(data)
+        if len(self.frames) >= self.max_frames:
+            frames_to_emit = self.frames[:]
+            self.frames = []
+            return frames_to_emit
+
+        return None
+
+    async def _flush(self):
+        frames = self.frames[:]
+        self.frames = []
+        if frames:
+            await self.emit(frames)
+
+
+AudioChunkerAutoProcessor.register("frames", AudioChunkerFramesProcessor)

server/reflector/processors/audio_chunker_silero.py

@@ -0,0 +1,298 @@
+from typing import Optional
+
+import av
+import numpy as np
+import torch
+from silero_vad import VADIterator, load_silero_vad
+
+from reflector.processors.audio_chunker import AudioChunkerProcessor
+from reflector.processors.audio_chunker_auto import AudioChunkerAutoProcessor
+
+
+class AudioChunkerSileroProcessor(AudioChunkerProcessor):
+    """
+    Assemble audio frames into chunks with VAD-based speech detection using Silero VAD
+    """
+
+    def __init__(
+        self,
+        block_frames=256,
+        max_frames=1024,
+        use_onnx=True,
+        min_frames=2,
+        **kwargs,
+    ):
+        super().__init__(**kwargs)
+        self.block_frames = block_frames
+        self.max_frames = max_frames
+        self.min_frames = min_frames
+
+        # Initialize Silero VAD
+        self._init_vad(use_onnx)
+
+    def _init_vad(self, use_onnx=False):
+        """Initialize Silero VAD model"""
+        try:
+            torch.set_num_threads(1)
+            self.vad_model = load_silero_vad(onnx=use_onnx)
+            self.vad_iterator = VADIterator(self.vad_model, sampling_rate=16000)
+            self.logger.info("Silero VAD initialized successfully")
+
+        except Exception as e:
+            self.logger.error(f"Failed to initialize Silero VAD: {e}")
+            self.vad_model = None
+            self.vad_iterator = None
+
+    async def _chunk(self, data: av.AudioFrame) -> Optional[list[av.AudioFrame]]:
+        """Process audio frame and return chunk when ready"""
+        self.frames.append(data)
+
+        # Check for speech segments every 32 frames (~1 second)
+        if len(self.frames) >= 32 and len(self.frames) % 32 == 0:
+            return await self._process_block()
+
+        # Safety fallback - emit if we hit max frames
+        elif len(self.frames) >= self.max_frames:
+            self.logger.warning(
+                f"AudioChunkerSileroProcessor: Reached max frames ({self.max_frames}), "
+                f"emitting first {self.max_frames // 2} frames"
+            )
+            frames_to_emit = self.frames[: self.max_frames // 2]
+            self.frames = self.frames[self.max_frames // 2 :]
+            if len(frames_to_emit) >= self.min_frames:
+                return frames_to_emit
+            else:
+                self.logger.debug(
+                    f"Ignoring fallback segment with {len(frames_to_emit)} frames "
+                    f"(< {self.min_frames} minimum)"
+                )
+
+        return None
+
+    async def _process_block(self) -> Optional[list[av.AudioFrame]]:
+        # Need at least 32 frames for VAD detection (~1 second)
+        if len(self.frames) < 32 or self.vad_iterator is None:
+            return None
+
+        # Processing block with current buffer size
+        print(f"Processing block: {len(self.frames)} frames in buffer")
+
+        try:
+            # Convert frames to numpy array for VAD
+            audio_array = self._frames_to_numpy(self.frames)
+
+            if audio_array is None:
+                # Fallback: emit all frames if conversion failed
+                frames_to_emit = self.frames[:]
+                self.frames = []
+                if len(frames_to_emit) >= self.min_frames:
+                    return frames_to_emit
+                else:
+                    self.logger.debug(
+                        f"Ignoring conversion-failed segment with {len(frames_to_emit)} frames "
+                        f"(< {self.min_frames} minimum)"
+                    )
+                return None
+
+            # Find complete speech segments in the buffer
+            speech_end_frame = self._find_speech_segment_end(audio_array)
+
+            if speech_end_frame is None or speech_end_frame <= 0:
+                # No speech found but buffer is getting large
+                if len(self.frames) > 512:
+                    # Check if it's all silence and can be discarded
+                    # No speech segment found, buffer at {len(self.frames)} frames
+
+                    # Could emit silence or discard old frames here
+                    # For now, keep first 256 frames and discard older silence
+                    if len(self.frames) > 768:
+                        self.logger.debug(
+                            f"Discarding {len(self.frames) - 256} old frames (likely silence)"
+                        )
+                        self.frames = self.frames[-256:]
+                return None
+
+            # Calculate segment timing information
+            frames_to_emit = self.frames[:speech_end_frame]
+
+            # Get timing from av.AudioFrame
+            if frames_to_emit:
+                first_frame = frames_to_emit[0]
+                last_frame = frames_to_emit[-1]
+                sample_rate = first_frame.sample_rate
+
+                # Calculate duration
+                total_samples = sum(f.samples for f in frames_to_emit)
+                duration_seconds = total_samples / sample_rate if sample_rate > 0 else 0
+
+                # Get timestamps if available
+                start_time = (
+                    first_frame.pts * first_frame.time_base if first_frame.pts else 0
+                )
+                end_time = (
+                    last_frame.pts * last_frame.time_base if last_frame.pts else 0
+                )
+
+                # Convert to HH:MM:SS format for logging
+                def format_time(seconds):
+                    if not seconds:
+                        return "00:00:00"
+                    total_seconds = int(float(seconds))
+                    hours = total_seconds // 3600
+                    minutes = (total_seconds % 3600) // 60
+                    secs = total_seconds % 60
+                    return f"{hours:02d}:{minutes:02d}:{secs:02d}"
+
+                start_formatted = format_time(start_time)
+                end_formatted = format_time(end_time)
+
+                # Keep remaining frames for next processing
+                remaining_after = len(self.frames) - speech_end_frame
+
+                # Single structured log line
+                self.logger.info(
+                    "Speech segment found",
+                    start=start_formatted,
+                    end=end_formatted,
+                    frames=speech_end_frame,
+                    duration=round(duration_seconds, 2),
+                    buffer_before=len(self.frames),
+                    remaining=remaining_after,
+                )
+
+            # Keep remaining frames for next processing
+            self.frames = self.frames[speech_end_frame:]
+
+            # Filter out segments with too few frames
+            if len(frames_to_emit) >= self.min_frames:
+                return frames_to_emit
+            else:
+                self.logger.debug(
+                    f"Ignoring segment with {len(frames_to_emit)} frames "
+                    f"(< {self.min_frames} minimum)"
+                )
+
+        except Exception as e:
+            self.logger.error(f"Error in VAD processing: {e}")
+            # Fallback to simple chunking
+            if len(self.frames) >= self.block_frames:
+                frames_to_emit = self.frames[: self.block_frames]
+                self.frames = self.frames[self.block_frames :]
+                if len(frames_to_emit) >= self.min_frames:
+                    return frames_to_emit
+                else:
+                    self.logger.debug(
+                        f"Ignoring exception-fallback segment with {len(frames_to_emit)} frames "
+                        f"(< {self.min_frames} minimum)"
+                    )
+
+        return None
+
+    def _frames_to_numpy(self, frames: list[av.AudioFrame]) -> Optional[np.ndarray]:
+        """Convert av.AudioFrame list to numpy array for VAD processing"""
+        if not frames:
+            return None
+
+        try:
+            audio_data = []
+            for frame in frames:
+                frame_array = frame.to_ndarray()
+
+                if len(frame_array.shape) == 2:
+                    frame_array = frame_array.flatten()
+
+                audio_data.append(frame_array)
+
+            if not audio_data:
+                return None
+
+            combined_audio = np.concatenate(audio_data)
+
+            # Ensure float32 format
+            if combined_audio.dtype == np.int16:
+                # Normalize int16 audio to float32 in range [-1.0, 1.0]
+                combined_audio = combined_audio.astype(np.float32) / 32768.0
+            elif combined_audio.dtype != np.float32:
+                combined_audio = combined_audio.astype(np.float32)
+
+            return combined_audio
+
+        except Exception as e:
+            self.logger.error(f"Error converting frames to numpy: {e}")
+
+        return None
+
+    def _find_speech_segment_end(self, audio_array: np.ndarray) -> Optional[int]:
+        """Find complete speech segments and return frame index at segment end"""
+        if self.vad_iterator is None or len(audio_array) == 0:
+            return None
+
+        try:
+            # Process audio in 512-sample windows for VAD
+            window_size = 512
+            min_silence_windows = 3  # Require 3 windows of silence after speech
+
+            # Track speech state
+            in_speech = False
+            speech_start = None
+            speech_end = None
+            silence_count = 0
+
+            for i in range(0, len(audio_array), window_size):
+                chunk = audio_array[i : i + window_size]
+                if len(chunk) < window_size:
+                    chunk = np.pad(chunk, (0, window_size - len(chunk)))
+
+                # Detect if this window has speech
+                speech_dict = self.vad_iterator(chunk, return_seconds=True)
+
+                # VADIterator returns dict with 'start' and 'end' when speech segments are detected
+                if speech_dict:
+                    if not in_speech:
+                        # Speech started
+                        speech_start = i
+                        in_speech = True
+                        # Debug: print(f"Speech START at sample {i}, VAD: {speech_dict}")
+                    silence_count = 0  # Reset silence counter
+                    continue
+
+                if not in_speech:
+                    continue
+
+                # We're in speech but found silence
+                silence_count += 1
+                if silence_count < min_silence_windows:
+                    continue
+
+                # Found end of speech segment
+                speech_end = i - (min_silence_windows - 1) * window_size
+                # Debug: print(f"Speech END at sample {speech_end}")
+
+                # Convert sample position to frame index
+                samples_per_frame = self.frames[0].samples if self.frames else 1024
+                frame_index = speech_end // samples_per_frame
+
+                # Ensure we don't exceed buffer
+                frame_index = min(frame_index, len(self.frames))
+                return frame_index
+
+            return None
+
+        except Exception as e:
+            self.logger.error(f"Error finding speech segment: {e}")
+            return None
+
+    async def _flush(self):
+        frames = self.frames[:]
+        self.frames = []
+        if frames:
+            if len(frames) >= self.min_frames:
+                await self.emit(frames)
+            else:
+                self.logger.debug(
+                    f"Ignoring flush segment with {len(frames)} frames "
+                    f"(< {self.min_frames} minimum)"
+                )
+
+
+AudioChunkerAutoProcessor.register("silero", AudioChunkerSileroProcessor)

server/reflector/processors/audio_downscale.py

@@ -0,0 +1,60 @@
+from typing import Optional
+
+import av
+from av.audio.resampler import AudioResampler
+
+from reflector.processors.base import Processor
+
+
+def copy_frame(frame: av.AudioFrame) -> av.AudioFrame:
+    frame_copy = frame.from_ndarray(
+        frame.to_ndarray(),
+        format=frame.format.name,
+        layout=frame.layout.name,
+    )
+    frame_copy.sample_rate = frame.sample_rate
+    frame_copy.pts = frame.pts
+    frame_copy.time_base = frame.time_base
+    return frame_copy
+
+
+class AudioDownscaleProcessor(Processor):
+    """
+    Downscale audio frames to 16kHz mono format
+    """
+
+    INPUT_TYPE = av.AudioFrame
+    OUTPUT_TYPE = av.AudioFrame
+
+    def __init__(self, target_rate: int = 16000, target_layout: str = "mono", **kwargs):
+        super().__init__(**kwargs)
+        self.target_rate = target_rate
+        self.target_layout = target_layout
+        self.resampler: Optional[AudioResampler] = None
+        self.needs_resampling: Optional[bool] = None
+
+    async def _push(self, data: av.AudioFrame):
+        if self.needs_resampling is None:
+            self.needs_resampling = (
+                data.sample_rate != self.target_rate
+                or data.layout.name != self.target_layout
+            )
+
+            if self.needs_resampling:
+                self.resampler = AudioResampler(
+                    format="s16", layout=self.target_layout, rate=self.target_rate
+                )
+
+        if not self.needs_resampling or not self.resampler:
+            await self.emit(data)
+            return
+
+        resampled_frames = self.resampler.resample(copy_frame(data))
+        for resampled_frame in resampled_frames:
+            await self.emit(resampled_frame)
+
+    async def _flush(self):
+        if self.needs_resampling and self.resampler:
+            final_frames = self.resampler.resample(None)
+            for frame in final_frames:
+                await self.emit(frame)

server/reflector/processors/audio_merge.py

@@ -3,24 +3,11 @@ from time import monotonic_ns
 from uuid import uuid4
 
 import av
-from av.audio.resampler import AudioResampler
 
 from reflector.processors.base import Processor
 from reflector.processors.types import AudioFile
 
 
-def copy_frame(frame: av.AudioFrame) -> av.AudioFrame:
-    frame_copy = frame.from_ndarray(
-        frame.to_ndarray(),
-        format=frame.format.name,
-        layout=frame.layout.name,
-    )
-    frame_copy.sample_rate = frame.sample_rate
-    frame_copy.pts = frame.pts
-    frame_copy.time_base = frame.time_base
-    return frame_copy
-
-
 class AudioMergeProcessor(Processor):
     """
     Merge audio frame into a single file
@@ -29,9 +16,8 @@ class AudioMergeProcessor(Processor):
     INPUT_TYPE = list[av.AudioFrame]
     OUTPUT_TYPE = AudioFile
 
-    def __init__(self, downsample_to_16k_mono: bool = True, **kwargs):
+    def __init__(self, **kwargs):
         super().__init__(**kwargs)
-        self.downsample_to_16k_mono = downsample_to_16k_mono
 
     async def _push(self, data: list[av.AudioFrame]):
         if not data:
@@ -39,72 +25,27 @@ class AudioMergeProcessor(Processor):
 
         # get audio information from first frame
         frame = data[0]
-        original_channels = len(frame.layout.channels)
-        original_sample_rate = frame.sample_rate
-        original_sample_width = frame.format.bytes
-
-        # determine if we need processing
-        needs_processing = self.downsample_to_16k_mono and (
-            original_sample_rate != 16000 or original_channels != 1
-        )
-
-        # determine output parameters
-        if self.downsample_to_16k_mono:
-            output_sample_rate = 16000
-            output_channels = 1
-            output_sample_width = 2  # 16-bit = 2 bytes
-        else:
-            output_sample_rate = original_sample_rate
-            output_channels = original_channels
-            output_sample_width = original_sample_width
+        output_channels = len(frame.layout.channels)
+        output_sample_rate = frame.sample_rate
+        output_sample_width = frame.format.bytes
 
         # create audio file
         uu = uuid4().hex
         fd = io.BytesIO()
 
-        if needs_processing:
-            # Process with PyAV resampler
-            out_container = av.open(fd, "w", format="wav")
-            out_stream = out_container.add_stream("pcm_s16le", rate=16000)
-            out_stream.layout = "mono"
+        # Use PyAV to write frames
+        out_container = av.open(fd, "w", format="wav")
+        out_stream = out_container.add_stream("pcm_s16le", rate=output_sample_rate)
+        out_stream.layout = frame.layout.name
 
-            # Create resampler if needed
-            resampler = None
-            if original_sample_rate != 16000 or original_channels != 1:
-                resampler = AudioResampler(format="s16", layout="mono", rate=16000)
-
-            for frame in data:
-                if resampler:
-                    # Resample and convert to mono
-                    # XXX for an unknown reason, if we don't use a copy of the frame, we get
-                    # Invalid Argumment from resample. Debugging indicate that when a previous processor
-                    # already used the frame (like AudioFileWriter), it make it invalid argument here.
-                    resampled_frames = resampler.resample(copy_frame(frame))
-                    for resampled_frame in resampled_frames:
-                        for packet in out_stream.encode(resampled_frame):
-                            out_container.mux(packet)
-                else:
-                    # Direct encoding without resampling
-                    for packet in out_stream.encode(frame):
-                        out_container.mux(packet)
-
-            # Flush the encoder
-            for packet in out_stream.encode(None):
+        for frame in data:
+            for packet in out_stream.encode(frame):
                 out_container.mux(packet)
-            out_container.close()
-        else:
-            # Use PyAV for original frames (no processing needed)
-            out_container = av.open(fd, "w", format="wav")
-            out_stream = out_container.add_stream("pcm_s16le", rate=output_sample_rate)
-            out_stream.layout = "mono" if output_channels == 1 else frame.layout
 
-            for frame in data:
-                for packet in out_stream.encode(frame):
-                    out_container.mux(packet)
-
-            for packet in out_stream.encode(None):
-                out_container.mux(packet)
-            out_container.close()
+        # Flush the encoder
+        for packet in out_stream.encode(None):
+            out_container.mux(packet)
+        out_container.close()
 
         fd.seek(0)
 

server/reflector/processors/audio_transcript_modal.py

@@ -12,9 +12,6 @@ API will be a POST request to TRANSCRIPT_URL:
 
 """
 
-from typing import List
-
-import aiohttp
 from openai import AsyncOpenAI
 
 from reflector.processors.audio_transcript import AudioTranscriptProcessor
@@ -25,7 +22,9 @@ from reflector.settings import settings
 
 class AudioTranscriptModalProcessor(AudioTranscriptProcessor):
     def __init__(
-        self, modal_api_key: str | None = None, batch_enabled: bool = True, **kwargs
+        self,
+        modal_api_key: str | None = None,
+        **kwargs,
     ):
         super().__init__()
         if not settings.TRANSCRIPT_URL:
@@ -35,126 +34,6 @@ class AudioTranscriptModalProcessor(AudioTranscriptProcessor):
         self.transcript_url = settings.TRANSCRIPT_URL + "/v1"
         self.timeout = settings.TRANSCRIPT_TIMEOUT
         self.modal_api_key = modal_api_key
-        self.max_batch_duration = 10.0
-        self.max_batch_files = 15
-        self.batch_enabled = batch_enabled
-        self.pending_files: List[AudioFile] = []  # Files waiting to be processed
-
-    @classmethod
-    def _calculate_duration(cls, audio_file: AudioFile) -> float:
-        """Calculate audio duration in seconds from AudioFile metadata"""
-        # Duration = total_samples / sample_rate
-        # We need to estimate total samples from the file data
-        import wave
-
-        try:
-            # Try to read as WAV file to get duration
-            audio_file.fd.seek(0)
-            with wave.open(audio_file.fd, "rb") as wav_file:
-                frames = wav_file.getnframes()
-                sample_rate = wav_file.getframerate()
-                duration = frames / sample_rate
-                return duration
-        except Exception:
-            # Fallback: estimate from file size and audio parameters
-            audio_file.fd.seek(0, 2)  # Seek to end
-            file_size = audio_file.fd.tell()
-            audio_file.fd.seek(0)  # Reset to beginning
-
-            # Estimate: file_size / (sample_rate * channels * sample_width)
-            bytes_per_second = (
-                audio_file.sample_rate
-                * audio_file.channels
-                * (audio_file.sample_width // 8)
-            )
-            estimated_duration = (
-                file_size / bytes_per_second if bytes_per_second > 0 else 0
-            )
-            return max(0, estimated_duration)
-
-    def _create_batches(self, audio_files: List[AudioFile]) -> List[List[AudioFile]]:
-        """Group audio files into batches with maximum 30s total duration"""
-        batches = []
-        current_batch = []
-        current_duration = 0.0
-
-        for audio_file in audio_files:
-            duration = self._calculate_duration(audio_file)
-
-            # If adding this file exceeds max duration, start a new batch
-            if current_duration + duration > self.max_batch_duration and current_batch:
-                batches.append(current_batch)
-                current_batch = [audio_file]
-                current_duration = duration
-            else:
-                current_batch.append(audio_file)
-                current_duration += duration
-
-        # Add the last batch if not empty
-        if current_batch:
-            batches.append(current_batch)
-
-        return batches
-
-    async def _transcript_batch(self, audio_files: List[AudioFile]) -> List[Transcript]:
-        """Transcribe a batch of audio files using the parakeet backend"""
-        if not audio_files:
-            return []
-
-        self.logger.debug(f"Batch transcribing {len(audio_files)} files")
-
-        # Prepare form data for batch request
-        data = aiohttp.FormData()
-        data.add_field("language", self.get_pref("audio:source_language", "en"))
-        data.add_field("batch", "true")
-
-        for i, audio_file in enumerate(audio_files):
-            audio_file.fd.seek(0)
-            data.add_field(
-                "files",
-                audio_file.fd,
-                filename=f"{audio_file.name}",
-                content_type="audio/wav",
-            )
-
-        # Make batch request
-        headers = {"Authorization": f"Bearer {self.modal_api_key}"}
-
-        async with aiohttp.ClientSession(
-            timeout=aiohttp.ClientTimeout(total=self.timeout)
-        ) as session:
-            async with session.post(
-                f"{self.transcript_url}/audio/transcriptions",
-                data=data,
-                headers=headers,
-            ) as response:
-                if response.status != 200:
-                    error_text = await response.text()
-                    raise Exception(
-                        f"Batch transcription failed: {response.status} {error_text}"
-                    )
-
-                result = await response.json()
-
-        # Process batch results
-        transcripts = []
-        results = result.get("results", [])
-
-        for i, (audio_file, file_result) in enumerate(zip(audio_files, results)):
-            transcript = Transcript(
-                words=[
-                    Word(
-                        text=word_info["word"],
-                        start=word_info["start"],
-                        end=word_info["end"],
-                    )
-                    for word_info in file_result.get("words", [])
-                ]
-            )
-            transcript.add_offset(audio_file.timestamp)
-            transcripts.append(transcript)
-
-        return transcripts
 
     async def _transcript(self, data: AudioFile):
         async with AsyncOpenAI(
@@ -187,96 +66,5 @@ class AudioTranscriptModalProcessor(AudioTranscriptProcessor):
 
         return transcript
 
-    async def transcript_multiple(
-        self, audio_files: List[AudioFile]
-    ) -> List[Transcript]:
-        """Transcribe multiple audio files using batching"""
-        if len(audio_files) == 1:
-            # Single file, use existing method
-            return [await self._transcript(audio_files[0])]
-
-        # Create batches with max 30s duration each
-        batches = self._create_batches(audio_files)
-
-        self.logger.debug(
-            f"Processing {len(audio_files)} files in {len(batches)} batches"
-        )
-
-        # Process all batches concurrently
-        all_transcripts = []
-
-        for batch in batches:
-            batch_transcripts = await self._transcript_batch(batch)
-            all_transcripts.extend(batch_transcripts)
-
-        return all_transcripts
-
-    async def _push(self, data: AudioFile):
-        """Override _push to support batching"""
-        if not self.batch_enabled:
-            # Use parent implementation for single file processing
-            return await super()._push(data)
-
-        # Add file to pending batch
-        self.pending_files.append(data)
-        self.logger.debug(
-            f"Added file to batch: {data.name}, batch size: {len(self.pending_files)}"
-        )
-
-        # Calculate total duration of pending files
-        total_duration = sum(self._calculate_duration(f) for f in self.pending_files)
-
-        # Process batch if it reaches max duration or has multiple files ready for optimization
-        should_process_batch = (
-            total_duration >= self.max_batch_duration
-            or len(self.pending_files) >= self.max_batch_files
-        )
-
-        if should_process_batch:
-            await self._process_pending_batch()
-
-    async def _process_pending_batch(self):
-        """Process all pending files as batches"""
-        if not self.pending_files:
-            return
-
-        self.logger.debug(f"Processing batch of {len(self.pending_files)} files")
-
-        try:
-            # Create batches respecting duration limit
-            batches = self._create_batches(self.pending_files)
-
-            # Process each batch
-            for batch in batches:
-                self.m_transcript_call.inc()
-                try:
-                    with self.m_transcript.time():
-                        # Use batch transcription
-                        transcripts = await self._transcript_batch(batch)
-
-                    self.m_transcript_success.inc()
-
-                    # Emit each transcript
-                    for transcript in transcripts:
-                        if transcript:
-                            await self.emit(transcript)
-
-                except Exception:
-                    self.m_transcript_failure.inc()
-                    raise
-                finally:
-                    # Release audio files
-                    for audio_file in batch:
-                        audio_file.release()
-
-        finally:
-            # Clear pending files
-            self.pending_files.clear()
-
-    async def _flush(self):
-        """Process any remaining files when flushing"""
-        await self._process_pending_batch()
-        await super()._flush()
-
 
 AudioTranscriptAutoProcessor.register("modal", AudioTranscriptModalProcessor)

server/reflector/processors/file_diarization_modal.py

@@ -47,6 +47,7 @@ class FileDiarizationModalProcessor(FileDiarizationProcessor):
                     "audio_file_url": data.audio_url,
                     "timestamp": 0,
                 },
+                follow_redirects=True,
             )
             response.raise_for_status()
             diarization_data = response.json()["diarization"]

server/reflector/processors/file_transcript_modal.py

@@ -54,6 +54,7 @@ class FileTranscriptModalProcessor(FileTranscriptProcessor):
                     "language": data.language,
                     "batch": True,
                 },
+                follow_redirects=True,
             )
             response.raise_for_status()
             result = response.json()
@@ -67,6 +68,9 @@ class FileTranscriptModalProcessor(FileTranscriptProcessor):
             for word_info in result.get("words", [])
         ]
 
+        # words come not in order
+        words.sort(key=lambda w: w.start)
+
         return Transcript(words=words)
 
 

server/reflector/settings.py

@@ -1,5 +1,8 @@
+from pydantic.types import PositiveInt
 from pydantic_settings import BaseSettings, SettingsConfigDict
 
+from reflector.utils.string import NonEmptyString
+
 
 class Settings(BaseSettings):
     model_config = SettingsConfigDict(
@@ -21,6 +24,10 @@ class Settings(BaseSettings):
     # local data directory
     DATA_DIR: str = "./data"
 
+    # Audio Chunking
+    # backends: silero, frames
+    AUDIO_CHUNKER_BACKEND: str = "frames"
+
     # Audio Transcription
     # backends: whisper, modal
     TRANSCRIPT_BACKEND: str = "whisper"
@@ -86,9 +93,8 @@ class Settings(BaseSettings):
     AUTH_JWT_PUBLIC_KEY: str | None = "authentik.monadical.com_public.pem"
     AUTH_JWT_AUDIENCE: str | None = None
 
-    # API public mode
-    # if set, all anonymous record will be public
     PUBLIC_MODE: bool = False
+    PUBLIC_DATA_RETENTION_DAYS: PositiveInt = 7
 
     # Min transcript length to generate topic + summary
     MIN_TRANSCRIPT_LENGTH: int = 750
@@ -116,13 +122,24 @@ class Settings(BaseSettings):
 
     # Whereby integration
     WHEREBY_API_URL: str = "https://api.whereby.dev/v1"
-    WHEREBY_API_KEY: str | None = None
+    WHEREBY_API_KEY: NonEmptyString | None = None
+
+    # Jibri integration
+    JIBRI_RECORDINGS_PATH: str = "/recordings"
     WHEREBY_WEBHOOK_SECRET: str | None = None
     AWS_WHEREBY_ACCESS_KEY_ID: str | None = None
     AWS_WHEREBY_ACCESS_KEY_SECRET: str | None = None
     AWS_PROCESS_RECORDING_QUEUE_URL: str | None = None
     SQS_POLLING_TIMEOUT_SECONDS: int = 60
 
+    # Jitsi Meet
+    JITSI_DOMAIN: str = "meet.jit.si"
+    JITSI_JWT_SECRET: str | None = None
+    JITSI_WEBHOOK_SECRET: str | None = None
+    JITSI_APP_ID: str = "reflector"
+    JITSI_JWT_ISSUER: str = "reflector"
+    JITSI_JWT_AUDIENCE: str = "jitsi"
+
     # Zulip integration
     ZULIP_REALM: str | None = None
     ZULIP_API_KEY: str | None = None

server/reflector/tools/cleanup_old_data.py

@@ -0,0 +1,72 @@
+#!/usr/bin/env python
+"""
+Manual cleanup tool for old public data.
+Uses the same implementation as the Celery worker task.
+"""
+
+import argparse
+import asyncio
+import sys
+
+import structlog
+
+from reflector.settings import settings
+from reflector.worker.cleanup import _cleanup_old_public_data
+
+logger = structlog.get_logger(__name__)
+
+
+async def cleanup_old_data(days: int = 7):
+    logger.info(
+        "Starting manual cleanup",
+        retention_days=days,
+        public_mode=settings.PUBLIC_MODE,
+    )
+
+    if not settings.PUBLIC_MODE:
+        logger.critical(
+            "WARNING: PUBLIC_MODE is False. "
+            "This tool is intended for public instances only."
+        )
+        raise Exception("Tool intended for public instances only")
+
+    result = await _cleanup_old_public_data(days=days)
+
+    if result:
+        logger.info(
+            "Cleanup completed",
+            transcripts_deleted=result.get("transcripts_deleted", 0),
+            meetings_deleted=result.get("meetings_deleted", 0),
+            recordings_deleted=result.get("recordings_deleted", 0),
+            errors_count=len(result.get("errors", [])),
+        )
+        if result.get("errors"):
+            logger.warning(
+                "Errors encountered during cleanup:", errors=result["errors"][:10]
+            )
+    else:
+        logger.info("Cleanup skipped or completed without results")
+
+
+def main():
+    parser = argparse.ArgumentParser(
+        description="Clean up old transcripts and meetings"
+    )
+    parser.add_argument(
+        "--days",
+        type=int,
+        default=7,
+        help="Number of days to keep data (default: 7)",
+    )
+
+    args = parser.parse_args()
+
+    if args.days < 1:
+        logger.error("Days must be at least 1")
+        sys.exit(1)
+
+    asyncio.run(cleanup_old_data(days=args.days))
+
+
+if __name__ == "__main__":
+    main()

server/reflector/tools/process.py

@@ -1,292 +1,204 @@
 """
 Process audio file with diarization support
-===========================================
-
-Extended version of process.py that includes speaker diarization.
-This tool processes audio files locally without requiring the full server infrastructure.
 """
 
+import argparse
 import asyncio
-import tempfile
-import uuid
+import json
+import shutil
+import sys
+import time
 from pathlib import Path
-from typing import List
-
-import av
+from typing import Any, Dict, List, Literal
 
+from reflector.db.transcripts import SourceKind, TranscriptTopic, transcripts_controller
 from reflector.logger import logger
-from reflector.processors import (
-    AudioChunkerProcessor,
-    AudioFileWriterProcessor,
-    AudioMergeProcessor,
-    AudioTranscriptAutoProcessor,
-    Pipeline,
-    PipelineEvent,
-    TranscriptFinalSummaryProcessor,
-    TranscriptFinalTitleProcessor,
-    TranscriptLinerProcessor,
-    TranscriptTopicDetectorProcessor,
-    TranscriptTranslatorAutoProcessor,
+from reflector.pipelines.main_file_pipeline import (
+    task_pipeline_file_process as task_pipeline_file_process,
 )
-from reflector.processors.base import BroadcastProcessor, Processor
-from reflector.processors.types import (
-    AudioDiarizationInput,
-    TitleSummary,
-    TitleSummaryWithId,
+from reflector.pipelines.main_live_pipeline import pipeline_post as live_pipeline_post
+from reflector.pipelines.main_live_pipeline import (
+    pipeline_process as live_pipeline_process,
 )
 
 
-class TopicCollectorProcessor(Processor):
-    """Collect topics for diarization"""
+def serialize_topics(topics: List[TranscriptTopic]) -> List[Dict[str, Any]]:
+    """Convert TranscriptTopic objects to JSON-serializable dicts"""
+    serialized = []
+    for topic in topics:
+        topic_dict = topic.model_dump()
+        serialized.append(topic_dict)
+    return serialized
 
-    INPUT_TYPE = TitleSummary
-    OUTPUT_TYPE = TitleSummary
 
-    def __init__(self, **kwargs):
-        super().__init__(**kwargs)
-        self.topics: List[TitleSummaryWithId] = []
-        self._topic_id = 0
+def debug_print_speakers(serialized_topics: List[Dict[str, Any]]) -> None:
+    """Print debug info about speakers found in topics"""
+    all_speakers = set()
+    for topic_dict in serialized_topics:
+        for word in topic_dict.get("words", []):
+            all_speakers.add(word.get("speaker", 0))
 
-    async def _push(self, data: TitleSummary):
-        # Convert to TitleSummaryWithId and collect
-        self._topic_id += 1
-        topic_with_id = TitleSummaryWithId(
-            id=str(self._topic_id),
-            title=data.title,
-            summary=data.summary,
-            timestamp=data.timestamp,
-            duration=data.duration,
-            transcript=data.transcript,
+    print(
+        f"Found {len(serialized_topics)} topics with speakers: {all_speakers}",
+        file=sys.stderr,
+    )
+
+
+TranscriptId = str
+
+
+# common interface for every flow: it needs an Entry in db with specific ceremony (file path + status + actual file in file system)
+# ideally we want to get rid of it at some point
+async def prepare_entry(
+    source_path: str,
+    source_language: str,
+    target_language: str,
+) -> TranscriptId:
+    file_path = Path(source_path)
+
+    transcript = await transcripts_controller.add(
+        file_path.name,
+        # note that the real file upload has SourceKind: LIVE for the reason of it's an error
+        source_kind=SourceKind.FILE,
+        source_language=source_language,
+        target_language=target_language,
+        user_id=None,
+    )
+
+    logger.info(
+        f"Created empty transcript {transcript.id} for file {file_path.name} because technically we need an empty transcript before we start transcript"
+    )
+
+    # pipelines expect files as upload.*
+
+    extension = file_path.suffix
+    upload_path = transcript.data_path / f"upload{extension}"
+    upload_path.parent.mkdir(parents=True, exist_ok=True)
+    shutil.copy2(source_path, upload_path)
+    logger.info(f"Copied {source_path} to {upload_path}")
+
+    # pipelines expect entity status "uploaded"
+    await transcripts_controller.update(transcript, {"status": "uploaded"})
+
+    return transcript.id
+
+
+# same reason as prepare_entry
+async def extract_result_from_entry(
+    transcript_id: TranscriptId, output_path: str
+) -> None:
+    post_final_transcript = await transcripts_controller.get_by_id(transcript_id)
+
+    # assert post_final_transcript.status == "ended"
+    # File pipeline doesn't set status to "ended", only live pipeline does https://github.com/Monadical-SAS/reflector/issues/582
+    topics = post_final_transcript.topics
+    if not topics:
+        raise RuntimeError(
+            f"No topics found for transcript {transcript_id} after processing"
         )
-        self.topics.append(topic_with_id)
 
-        # Pass through the original topic
-        await self.emit(data)
+    serialized_topics = serialize_topics(topics)
 
-    def get_topics(self) -> List[TitleSummaryWithId]:
-        return self.topics
+    if output_path:
+        # Write to JSON file
+        with open(output_path, "w") as f:
+            for topic_dict in serialized_topics:
+                json.dump(topic_dict, f)
+                f.write("\n")
+        print(f"Results written to {output_path}", file=sys.stderr)
+    else:
+        # Write to stdout as JSONL
+        for topic_dict in serialized_topics:
+            print(json.dumps(topic_dict))
+
+    debug_print_speakers(serialized_topics)
 
 
-async def process_audio_file(
-    filename,
-    event_callback,
-    only_transcript=False,
-    source_language="en",
-    target_language="en",
-    enable_diarization=True,
-    diarization_backend="pyannote",
+async def process_live_pipeline(
+    transcript_id: TranscriptId,
 ):
-    # Create temp file for audio if diarization is enabled
-    audio_temp_path = None
-    if enable_diarization:
-        audio_temp_file = tempfile.NamedTemporaryFile(suffix=".wav", delete=False)
-        audio_temp_path = audio_temp_file.name
-        audio_temp_file.close()
+    """Process transcript_id with transcription and diarization"""
 
-    # Create processor for collecting topics
-    topic_collector = TopicCollectorProcessor()
+    print(f"Processing transcript_id {transcript_id}...", file=sys.stderr)
+    await live_pipeline_process(transcript_id=transcript_id)
+    print(f"Processing complete for transcript {transcript_id}", file=sys.stderr)
 
-    # Build pipeline for audio processing
-    processors = []
+    pre_final_transcript = await transcripts_controller.get_by_id(transcript_id)
 
-    # Add audio file writer at the beginning if diarization is enabled
-    if enable_diarization:
-        processors.append(AudioFileWriterProcessor(audio_temp_path))
+    # assert documented behaviour: after process, the pipeline isn't ended. this is the reason of calling pipeline_post
+    assert pre_final_transcript.status != "ended"
 
-    # Add the rest of the processors
-    processors += [
-        AudioChunkerProcessor(),
-        AudioMergeProcessor(),
-        AudioTranscriptAutoProcessor.as_threaded(),
-        TranscriptLinerProcessor(),
-        TranscriptTranslatorAutoProcessor.as_threaded(),
-    ]
+    # at this point, diarization is running but we have no access to it. run diarization in parallel - one will hopefully win after polling
+    result = live_pipeline_post(transcript_id=transcript_id)
 
-    if not only_transcript:
-        processors += [
-            TranscriptTopicDetectorProcessor.as_threaded(),
-            # Collect topics for diarization
-            topic_collector,
-            BroadcastProcessor(
-                processors=[
-                    TranscriptFinalTitleProcessor.as_threaded(),
-                    TranscriptFinalSummaryProcessor.as_threaded(),
-                ],
-            ),
-        ]
-
-    # Create main pipeline
-    pipeline = Pipeline(*processors)
-    pipeline.set_pref("audio:source_language", source_language)
-    pipeline.set_pref("audio:target_language", target_language)
-    pipeline.describe()
-    pipeline.on(event_callback)
-
-    # Start processing audio
-    logger.info(f"Opening {filename}")
-    container = av.open(filename)
-    try:
-        logger.info("Start pushing audio into the pipeline")
-        for frame in container.decode(audio=0):
-            await pipeline.push(frame)
-    finally:
-        logger.info("Flushing the pipeline")
-        await pipeline.flush()
-
-    # Run diarization if enabled and we have topics
-    if enable_diarization and not only_transcript and audio_temp_path:
-        topics = topic_collector.get_topics()
-
-        if topics:
-            logger.info(f"Starting diarization with {len(topics)} topics")
-
-            try:
-                from reflector.processors import AudioDiarizationAutoProcessor
-
-                diarization_processor = AudioDiarizationAutoProcessor(
-                    name=diarization_backend
-                )
-
-                diarization_processor.set_pipeline(pipeline)
-
-                # For Modal backend, we need to upload the file to S3 first
-                if diarization_backend == "modal":
-                    from datetime import datetime
-
-                    from reflector.storage import get_transcripts_storage
-                    from reflector.utils.s3_temp_file import S3TemporaryFile
-
-                    storage = get_transcripts_storage()
-
-                    # Generate a unique filename in evaluation folder
-                    timestamp = datetime.utcnow().strftime("%Y%m%d_%H%M%S")
-                    audio_filename = f"evaluation/diarization_temp/{timestamp}_{uuid.uuid4().hex}.wav"
-
-                    # Use context manager for automatic cleanup
-                    async with S3TemporaryFile(storage, audio_filename) as s3_file:
-                        # Read and upload the audio file
-                        with open(audio_temp_path, "rb") as f:
-                            audio_data = f.read()
-
-                        audio_url = await s3_file.upload(audio_data)
-                        logger.info(f"Uploaded audio to S3: {audio_filename}")
-
-                        # Create diarization input with S3 URL
-                        diarization_input = AudioDiarizationInput(
-                            audio_url=audio_url, topics=topics
-                        )
-
-                        # Run diarization
-                        await diarization_processor.push(diarization_input)
-                        await diarization_processor.flush()
-
-                        logger.info("Diarization complete")
-                        # File will be automatically cleaned up when exiting the context
-                else:
-                    # For local backend, use local file path
-                    audio_url = audio_temp_path
-
-                    # Create diarization input
-                    diarization_input = AudioDiarizationInput(
-                        audio_url=audio_url, topics=topics
-                    )
-
-                    # Run diarization
-                    await diarization_processor.push(diarization_input)
-                    await diarization_processor.flush()
-
-                    logger.info("Diarization complete")
-
-            except ImportError as e:
-                logger.error(f"Failed to import diarization dependencies: {e}")
-                logger.error(
-                    "Install with: uv pip install pyannote.audio torch torchaudio"
-                )
-                logger.error(
-                    "And set HF_TOKEN environment variable for pyannote models"
-                )
-                raise SystemExit(1)
-            except Exception as e:
-                logger.error(f"Diarization failed: {e}")
-                raise SystemExit(1)
-        else:
-            logger.warning("Skipping diarization: no topics available")
-
-    # Clean up temp file
-    if audio_temp_path:
-        try:
-            Path(audio_temp_path).unlink()
-        except Exception as e:
-            logger.warning(f"Failed to clean up temp file {audio_temp_path}: {e}")
-
-    logger.info("All done!")
+    # result.ready() blocks even without await; it mutates result also
+    while not result.ready():
+        print(f"Status: {result.state}")
+        time.sleep(2)
 
 
 async def process_file_pipeline(
-    filename: str,
-    event_callback,
-    source_language="en",
-    target_language="en",
-    enable_diarization=True,
-    diarization_backend="modal",
+    transcript_id: TranscriptId,
 ):
     """Process audio/video file using the optimized file pipeline"""
+
+    # task_pipeline_file_process is a Celery task, need to use .delay() for async execution
+    result = task_pipeline_file_process.delay(transcript_id=transcript_id)
+
+    # Wait for the Celery task to complete
+    while not result.ready():
+        print(f"File pipeline status: {result.state}", file=sys.stderr)
+        time.sleep(2)
+
+    logger.info("File pipeline processing complete")
+
+
+async def process(
+    source_path: str,
+    source_language: str,
+    target_language: str,
+    pipeline: Literal["live", "file"],
+    output_path: str = None,
+):
+    from reflector.db import get_database
+
+    database = get_database()
+    # db connect is a part of ceremony
+    await database.connect()
+
     try:
-        from reflector.db import database
-        from reflector.db.transcripts import SourceKind, transcripts_controller
-        from reflector.pipelines.main_file_pipeline import PipelineMainFile
-
-        await database.connect()
-        try:
-            # Create a temporary transcript for processing
-            transcript = await transcripts_controller.add(
-                "",
-                source_kind=SourceKind.FILE,
-                source_language=source_language,
-                target_language=target_language,
-            )
-
-            # Process the file
-            pipeline = PipelineMainFile(transcript_id=transcript.id)
-            await pipeline.process(Path(filename))
-
-            logger.info("File pipeline processing complete")
-
-        finally:
-            await database.disconnect()
-    except ImportError as e:
-        logger.error(f"File pipeline not available: {e}")
-        logger.info("Falling back to stream pipeline")
-        # Fall back to stream pipeline
-        await process_audio_file(
-            filename,
-            event_callback,
-            only_transcript=False,
-            source_language=source_language,
-            target_language=target_language,
-            enable_diarization=enable_diarization,
-            diarization_backend=diarization_backend,
+        transcript_id = await prepare_entry(
+            source_path,
+            source_language,
+            target_language,
         )
 
+        pipeline_handlers = {
+            "live": process_live_pipeline,
+            "file": process_file_pipeline,
+        }
+
+        handler = pipeline_handlers.get(pipeline)
+        if not handler:
+            raise ValueError(f"Unknown pipeline type: {pipeline}")
+
+        await handler(transcript_id)
+
+        await extract_result_from_entry(transcript_id, output_path)
+    finally:
+        await database.disconnect()
+
 
 if __name__ == "__main__":
-    import argparse
-    import os
-
     parser = argparse.ArgumentParser(
-        description="Process audio files with optional speaker diarization"
+        description="Process audio files with speaker diarization"
     )
     parser.add_argument("source", help="Source file (mp3, wav, mp4...)")
     parser.add_argument(
-        "--stream",
-        action="store_true",
-        help="Use streaming pipeline (original frame-based processing)",
-    )
-    parser.add_argument(
-        "--only-transcript",
-        "-t",
-        action="store_true",
-        help="Only generate transcript without topics/summaries",
+        "--pipeline",
+        required=True,
+        choices=["live", "file"],
+        help="Pipeline type to use for processing (live: streaming/incremental, file: batch/parallel)",
     )
     parser.add_argument(
         "--source-language", default="en", help="Source language code (default: en)"
@@ -295,81 +207,14 @@ if __name__ == "__main__":
         "--target-language", default="en", help="Target language code (default: en)"
     )
     parser.add_argument("--output", "-o", help="Output file (output.jsonl)")
-    parser.add_argument(
-        "--enable-diarization",
-        "-d",
-        action="store_true",
-        help="Enable speaker diarization",
-    )
-    parser.add_argument(
-        "--diarization-backend",
-        default="pyannote",
-        choices=["pyannote", "modal"],
-        help="Diarization backend to use (default: pyannote)",
-    )
     args = parser.parse_args()
 
-    if "REDIS_HOST" not in os.environ:
-        os.environ["REDIS_HOST"] = "localhost"
-
-    output_fd = None
-    if args.output:
-        output_fd = open(args.output, "w")
-
-    async def event_callback(event: PipelineEvent):
-        processor = event.processor
-        data = event.data
-
-        # Ignore internal processors
-        if processor in (
-            "AudioChunkerProcessor",
-            "AudioMergeProcessor",
-            "AudioFileWriterProcessor",
-            "TopicCollectorProcessor",
-            "BroadcastProcessor",
-        ):
-            return
-
-        # If diarization is enabled, skip the original topic events from the pipeline
-        # The diarization processor will emit the same topics but with speaker info
-        if processor == "TranscriptTopicDetectorProcessor" and args.enable_diarization:
-            return
-
-        # Log all events
-        logger.info(f"Event: {processor} - {type(data).__name__}")
-
-        # Write to output
-        if output_fd:
-            output_fd.write(event.model_dump_json())
-            output_fd.write("\n")
-            output_fd.flush()
-
-    if args.stream:
-        # Use original streaming pipeline
-        asyncio.run(
-            process_audio_file(
-                args.source,
-                event_callback,
-                only_transcript=args.only_transcript,
-                source_language=args.source_language,
-                target_language=args.target_language,
-                enable_diarization=args.enable_diarization,
-                diarization_backend=args.diarization_backend,
-            )
+    asyncio.run(
+        process(
+            args.source,
+            args.source_language,
+            args.target_language,
+            args.pipeline,
+            args.output,
         )
-    else:
-        # Use optimized file pipeline (default)
-        asyncio.run(
-            process_file_pipeline(
-                args.source,
-                event_callback,
-                source_language=args.source_language,
-                target_language=args.target_language,
-                enable_diarization=args.enable_diarization,
-                diarization_backend=args.diarization_backend,
-            )
-        )
-
-    if output_fd:
-        output_fd.close()
-        logger.info(f"Output written to {args.output}")
+    )

server/reflector/tools/process_with_diarization.py

@@ -1,315 +0,0 @@
-"""
-@vibe-generated
-Process audio file with diarization support
-===========================================
-
-Extended version of process.py that includes speaker diarization.
-This tool processes audio files locally without requiring the full server infrastructure.
-"""
-
-import asyncio
-import tempfile
-import uuid
-from pathlib import Path
-from typing import List
-
-import av
-
-from reflector.logger import logger
-from reflector.processors import (
-    AudioChunkerProcessor,
-    AudioFileWriterProcessor,
-    AudioMergeProcessor,
-    AudioTranscriptAutoProcessor,
-    Pipeline,
-    PipelineEvent,
-    TranscriptFinalSummaryProcessor,
-    TranscriptFinalTitleProcessor,
-    TranscriptLinerProcessor,
-    TranscriptTopicDetectorProcessor,
-    TranscriptTranslatorAutoProcessor,
-)
-from reflector.processors.base import BroadcastProcessor, Processor
-from reflector.processors.types import (
-    AudioDiarizationInput,
-    TitleSummary,
-    TitleSummaryWithId,
-)
-
-
-class TopicCollectorProcessor(Processor):
-    """Collect topics for diarization"""
-
-    INPUT_TYPE = TitleSummary
-    OUTPUT_TYPE = TitleSummary
-
-    def __init__(self, **kwargs):
-        super().__init__(**kwargs)
-        self.topics: List[TitleSummaryWithId] = []
-        self._topic_id = 0
-
-    async def _push(self, data: TitleSummary):
-        # Convert to TitleSummaryWithId and collect
-        self._topic_id += 1
-        topic_with_id = TitleSummaryWithId(
-            id=str(self._topic_id),
-            title=data.title,
-            summary=data.summary,
-            timestamp=data.timestamp,
-            duration=data.duration,
-            transcript=data.transcript,
-        )
-        self.topics.append(topic_with_id)
-
-        # Pass through the original topic
-        await self.emit(data)
-
-    def get_topics(self) -> List[TitleSummaryWithId]:
-        return self.topics
-
-
-async def process_audio_file_with_diarization(
-    filename,
-    event_callback,
-    only_transcript=False,
-    source_language="en",
-    target_language="en",
-    enable_diarization=True,
-    diarization_backend="modal",
-):
-    # Create temp file for audio if diarization is enabled
-    audio_temp_path = None
-    if enable_diarization:
-        audio_temp_file = tempfile.NamedTemporaryFile(suffix=".wav", delete=False)
-        audio_temp_path = audio_temp_file.name
-        audio_temp_file.close()
-
-    # Create processor for collecting topics
-    topic_collector = TopicCollectorProcessor()
-
-    # Build pipeline for audio processing
-    processors = []
-
-    # Add audio file writer at the beginning if diarization is enabled
-    if enable_diarization:
-        processors.append(AudioFileWriterProcessor(audio_temp_path))
-
-    # Add the rest of the processors
-    processors += [
-        AudioChunkerProcessor(),
-        AudioMergeProcessor(),
-        AudioTranscriptAutoProcessor.as_threaded(),
-    ]
-
-    processors += [
-        TranscriptLinerProcessor(),
-        TranscriptTranslatorAutoProcessor.as_threaded(),
-    ]
-
-    if not only_transcript:
-        processors += [
-            TranscriptTopicDetectorProcessor.as_threaded(),
-            # Collect topics for diarization
-            topic_collector,
-            BroadcastProcessor(
-                processors=[
-                    TranscriptFinalTitleProcessor.as_threaded(),
-                    TranscriptFinalSummaryProcessor.as_threaded(),
-                ],
-            ),
-        ]
-
-    # Create main pipeline
-    pipeline = Pipeline(*processors)
-    pipeline.set_pref("audio:source_language", source_language)
-    pipeline.set_pref("audio:target_language", target_language)
-    pipeline.describe()
-    pipeline.on(event_callback)
-
-    # Start processing audio
-    logger.info(f"Opening {filename}")
-    container = av.open(filename)
-    try:
-        logger.info("Start pushing audio into the pipeline")
-        for frame in container.decode(audio=0):
-            await pipeline.push(frame)
-    finally:
-        logger.info("Flushing the pipeline")
-        await pipeline.flush()
-
-    # Run diarization if enabled and we have topics
-    if enable_diarization and not only_transcript and audio_temp_path:
-        topics = topic_collector.get_topics()
-
-        if topics:
-            logger.info(f"Starting diarization with {len(topics)} topics")
-
-            try:
-                from reflector.processors import AudioDiarizationAutoProcessor
-
-                diarization_processor = AudioDiarizationAutoProcessor(
-                    name=diarization_backend
-                )
-
-                diarization_processor.set_pipeline(pipeline)
-
-                # For Modal backend, we need to upload the file to S3 first
-                if diarization_backend == "modal":
-                    from datetime import datetime, timezone
-
-                    from reflector.storage import get_transcripts_storage
-                    from reflector.utils.s3_temp_file import S3TemporaryFile
-
-                    storage = get_transcripts_storage()
-
-                    # Generate a unique filename in evaluation folder
-                    timestamp = datetime.now(timezone.utc).strftime("%Y%m%d_%H%M%S")
-                    audio_filename = f"evaluation/diarization_temp/{timestamp}_{uuid.uuid4().hex}.wav"
-
-                    # Use context manager for automatic cleanup
-                    async with S3TemporaryFile(storage, audio_filename) as s3_file:
-                        # Read and upload the audio file
-                        with open(audio_temp_path, "rb") as f:
-                            audio_data = f.read()
-
-                        audio_url = await s3_file.upload(audio_data)
-                        logger.info(f"Uploaded audio to S3: {audio_filename}")
-
-                        # Create diarization input with S3 URL
-                        diarization_input = AudioDiarizationInput(
-                            audio_url=audio_url, topics=topics
-                        )
-
-                        # Run diarization
-                        await diarization_processor.push(diarization_input)
-                        await diarization_processor.flush()
-
-                        logger.info("Diarization complete")
-                        # File will be automatically cleaned up when exiting the context
-                else:
-                    # For local backend, use local file path
-                    audio_url = audio_temp_path
-
-                    # Create diarization input
-                    diarization_input = AudioDiarizationInput(
-                        audio_url=audio_url, topics=topics
-                    )
-
-                    # Run diarization
-                    await diarization_processor.push(diarization_input)
-                    await diarization_processor.flush()
-
-                    logger.info("Diarization complete")
-
-            except ImportError as e:
-                logger.error(f"Failed to import diarization dependencies: {e}")
-                logger.error(
-                    "Install with: uv pip install pyannote.audio torch torchaudio"
-                )
-                logger.error(
-                    "And set HF_TOKEN environment variable for pyannote models"
-                )
-                raise SystemExit(1)
-            except Exception as e:
-                logger.error(f"Diarization failed: {e}")
-                raise SystemExit(1)
-        else:
-            logger.warning("Skipping diarization: no topics available")
-
-    # Clean up temp file
-    if audio_temp_path:
-        try:
-            Path(audio_temp_path).unlink()
-        except Exception as e:
-            logger.warning(f"Failed to clean up temp file {audio_temp_path}: {e}")
-
-    logger.info("All done!")
-
-
-if __name__ == "__main__":
-    import argparse
-    import os
-
-    parser = argparse.ArgumentParser(
-        description="Process audio files with optional speaker diarization"
-    )
-    parser.add_argument("source", help="Source file (mp3, wav, mp4...)")
-    parser.add_argument(
-        "--only-transcript",
-        "-t",
-        action="store_true",
-        help="Only generate transcript without topics/summaries",
-    )
-    parser.add_argument(
-        "--source-language", default="en", help="Source language code (default: en)"
-    )
-    parser.add_argument(
-        "--target-language", default="en", help="Target language code (default: en)"
-    )
-    parser.add_argument("--output", "-o", help="Output file (output.jsonl)")
-    parser.add_argument(
-        "--enable-diarization",
-        "-d",
-        action="store_true",
-        help="Enable speaker diarization",
-    )
-    parser.add_argument(
-        "--diarization-backend",
-        default="modal",
-        choices=["modal"],
-        help="Diarization backend to use (default: modal)",
-    )
-    args = parser.parse_args()
-
-    # Set REDIS_HOST to localhost if not provided
-    if "REDIS_HOST" not in os.environ:
-        os.environ["REDIS_HOST"] = "localhost"
-        logger.info("REDIS_HOST not set, defaulting to localhost")
-
-    output_fd = None
-    if args.output:
-        output_fd = open(args.output, "w")
-
-    async def event_callback(event: PipelineEvent):
-        processor = event.processor
-        data = event.data
-
-        # Ignore internal processors
-        if processor in (
-            "AudioChunkerProcessor",
-            "AudioMergeProcessor",
-            "AudioFileWriterProcessor",
-            "TopicCollectorProcessor",
-            "BroadcastProcessor",
-        ):
-            return
-
-        # If diarization is enabled, skip the original topic events from the pipeline
-        # The diarization processor will emit the same topics but with speaker info
-        if processor == "TranscriptTopicDetectorProcessor" and args.enable_diarization:
-            return
-
-        # Log all events
-        logger.info(f"Event: {processor} - {type(data).__name__}")
-
-        # Write to output
-        if output_fd:
-            output_fd.write(event.model_dump_json())
-            output_fd.write("\n")
-            output_fd.flush()
-
-    asyncio.run(
-        process_audio_file_with_diarization(
-            args.source,
-            event_callback,
-            only_transcript=args.only_transcript,
-            source_language=args.source_language,
-            target_language=args.target_language,
-            enable_diarization=args.enable_diarization,
-            diarization_backend=args.diarization_backend,
-        )
-    )
-
-    if output_fd:
-        output_fd.close()
-        logger.info(f"Output written to {args.output}")

server/reflector/tools/runpipeline.py

@@ -53,7 +53,7 @@ async def run_single_processor(args):
     async def event_callback(event: PipelineEvent):
         processor = event.processor
         # ignore some processor
-        if processor in ("AudioChunkerProcessor", "AudioMergeProcessor"):
+        if processor in ("AudioChunkerAutoProcessor", "AudioMergeProcessor"):
             return
         print(f"Event: {event}")
         if output_fd:

server/reflector/tools/test_diarization.py

@@ -1,96 +0,0 @@
-#!/usr/bin/env python3
-"""
-@vibe-generated
-Test script for the diarization CLI tool
-=========================================
-
-This script helps test the diarization functionality with sample audio files.
-"""
-
-import asyncio
-import sys
-from pathlib import Path
-
-from reflector.logger import logger
-
-
-async def test_diarization(audio_file: str):
-    """Test the diarization functionality"""
-
-    # Import the processing function
-    from process_with_diarization import process_audio_file_with_diarization
-
-    # Collect events
-    events = []
-
-    async def event_callback(event):
-        events.append({"processor": event.processor, "data": event.data})
-        logger.info(f"Event from {event.processor}")
-
-    # Process the audio file
-    logger.info(f"Processing audio file: {audio_file}")
-
-    try:
-        await process_audio_file_with_diarization(
-            audio_file,
-            event_callback,
-            only_transcript=False,
-            source_language="en",
-            target_language="en",
-            enable_diarization=True,
-            diarization_backend="modal",
-        )
-
-        # Analyze results
-        logger.info(f"Processing complete. Received {len(events)} events")
-
-        # Look for diarization results
-        diarized_topics = []
-        for event in events:
-            if "TitleSummary" in event["processor"]:
-                # Check if words have speaker information
-                if hasattr(event["data"], "transcript") and event["data"].transcript:
-                    words = event["data"].transcript.words
-                    if words and hasattr(words[0], "speaker"):
-                        speakers = set(
-                            w.speaker for w in words if hasattr(w, "speaker")
-                        )
-                        logger.info(
-                            f"Found {len(speakers)} speakers in topic: {event['data'].title}"
-                        )
-                        diarized_topics.append(event["data"])
-
-        if diarized_topics:
-            logger.info(f"Successfully diarized {len(diarized_topics)} topics")
-
-            # Print sample output
-            sample_topic = diarized_topics[0]
-            logger.info("Sample diarized output:")
-            for i, word in enumerate(sample_topic.transcript.words[:10]):
-                logger.info(f"  Word {i}: '{word.text}' - Speaker {word.speaker}")
-        else:
-            logger.warning("No diarization results found in output")
-
-        return events
-
-    except Exception as e:
-        logger.error(f"Error during processing: {e}")
-        raise
-
-
-def main():
-    if len(sys.argv) < 2:
-        print("Usage: python test_diarization.py <audio_file>")
-        sys.exit(1)
-
-    audio_file = sys.argv[1]
-    if not Path(audio_file).exists():
-        print(f"Error: Audio file '{audio_file}' not found")
-        sys.exit(1)
-
-    # Run the test
-    asyncio.run(test_diarization(audio_file))
-
-
-if __name__ == "__main__":
-    main()

server/reflector/utils/string.py

@@ -0,0 +1,23 @@
+from typing import Annotated
+
+from pydantic import Field, TypeAdapter, constr
+
+NonEmptyStringBase = constr(min_length=1, strip_whitespace=False)
+NonEmptyString = Annotated[
+    NonEmptyStringBase,
+    Field(description="A non-empty string", min_length=1),
+]
+non_empty_string_adapter = TypeAdapter(NonEmptyString)
+
+
+def parse_non_empty_string(s: str, error: str | None = None) -> NonEmptyString:
+    try:
+        return non_empty_string_adapter.validate_python(s)
+    except Exception as e:
+        raise ValueError(f"{e}: {error}" if error else e) from e
+
+
+def try_parse_non_empty_string(s: str) -> NonEmptyString | None:
+    if not s:
+        return None
+    return parse_non_empty_string(s)

server/reflector/video_platforms/__init__.py

@@ -0,0 +1,17 @@
+# Video Platform Abstraction Layer
+"""
+This module provides an abstraction layer for different video conferencing platforms.
+It allows seamless switching between providers (Whereby, Daily.co, etc.) without
+changing the core application logic.
+"""
+
+from .base import MeetingData, VideoPlatformClient, VideoPlatformConfig
+from .registry import get_platform_client, register_platform
+
+__all__ = [
+    "VideoPlatformClient",
+    "VideoPlatformConfig",
+    "MeetingData",
+    "get_platform_client",
+    "register_platform",
+]

server/reflector/video_platforms/base.py

@@ -0,0 +1,82 @@
+from abc import ABC, abstractmethod
+from datetime import datetime
+from typing import Any, Dict, Optional
+
+from pydantic import BaseModel
+
+from reflector.db.rooms import Room
+
+
+class MeetingData(BaseModel):
+    """Standardized meeting data returned by all platforms."""
+
+    meeting_id: str
+    room_name: str
+    room_url: str
+    host_room_url: str
+    platform: str
+    extra_data: Dict[str, Any] = {}  # Platform-specific data
+
+
+class VideoPlatformConfig(BaseModel):
+    """Configuration for a video platform."""
+
+    api_key: str
+    webhook_secret: str
+    api_url: Optional[str] = None
+    subdomain: Optional[str] = None
+    s3_bucket: Optional[str] = None
+    s3_region: Optional[str] = None
+    aws_role_arn: Optional[str] = None
+    aws_access_key_id: Optional[str] = None
+    aws_access_key_secret: Optional[str] = None
+
+
+class VideoPlatformClient(ABC):
+    """Abstract base class for video platform integrations."""
+
+    PLATFORM_NAME: str = ""
+
+    def __init__(self, config: VideoPlatformConfig):
+        self.config = config
+
+    @abstractmethod
+    async def create_meeting(
+        self, room_name_prefix: str, end_date: datetime, room: Room
+    ) -> MeetingData:
+        """Create a new meeting room."""
+        pass
+
+    @abstractmethod
+    async def get_room_sessions(self, room_name: str) -> Dict[str, Any]:
+        """Get session information for a room."""
+        pass
+
+    @abstractmethod
+    async def delete_room(self, room_name: str) -> bool:
+        """Delete a room. Returns True if successful."""
+        pass
+
+    @abstractmethod
+    async def upload_logo(self, room_name: str, logo_path: str) -> bool:
+        """Upload a logo to the room. Returns True if successful."""
+        pass
+
+    @abstractmethod
+    def verify_webhook_signature(
+        self, body: bytes, signature: str, timestamp: Optional[str] = None
+    ) -> bool:
+        """Verify webhook signature for security."""
+        pass
+
+    def format_recording_config(self, room: Room) -> Dict[str, Any]:
+        """Format recording configuration for the platform.
+        Can be overridden by specific implementations."""
+        if room.recording_type == "cloud" and self.config.s3_bucket:
+            return {
+                "type": room.recording_type,
+                "bucket": self.config.s3_bucket,
+                "region": self.config.s3_region,
+                "trigger": room.recording_trigger,
+            }
+        return {"type": room.recording_type}

server/reflector/video_platforms/factory.py

@@ -0,0 +1,54 @@
+"""Factory for creating video platform clients based on configuration."""
+
+from typing import TYPE_CHECKING, Literal, Optional, overload
+
+from reflector.db.rooms import VideoPlatform
+from reflector.settings import settings
+
+from .base import VideoPlatformClient, VideoPlatformConfig
+from .registry import get_platform_client
+
+if TYPE_CHECKING:
+    from .jitsi import JitsiClient
+    from .whereby import WherebyClient
+
+
+def get_platform_config(platform: str) -> VideoPlatformConfig:
+    """Get configuration for a specific platform."""
+    if platform == VideoPlatform.WHEREBY:
+        return VideoPlatformConfig(
+            api_key=settings.WHEREBY_API_KEY or "",
+            webhook_secret=settings.WHEREBY_WEBHOOK_SECRET or "",
+            api_url=settings.WHEREBY_API_URL,
+            s3_bucket=settings.RECORDING_STORAGE_AWS_BUCKET_NAME,
+            aws_access_key_id=settings.AWS_WHEREBY_ACCESS_KEY_ID,
+            aws_access_key_secret=settings.AWS_WHEREBY_ACCESS_KEY_SECRET,
+        )
+    elif platform == VideoPlatform.JITSI:
+        return VideoPlatformConfig(
+            api_key="",  # Jitsi uses JWT, no API key
+            webhook_secret=settings.JITSI_WEBHOOK_SECRET or "",
+            api_url=f"https://{settings.JITSI_DOMAIN}",
+        )
+    else:
+        raise ValueError(f"Unknown platform: {platform}")
+
+
+@overload
+def create_platform_client(platform: Literal["jitsi"]) -> "JitsiClient": ...
+
+
+@overload
+def create_platform_client(platform: Literal["whereby"]) -> "WherebyClient": ...
+
+
+def create_platform_client(platform: str) -> VideoPlatformClient:
+    """Create a video platform client instance."""
+    config = get_platform_config(platform)
+    return get_platform_client(platform, config)
+
+
+def get_platform_for_room(room_id: Optional[str] = None) -> str:
+    """Determine which platform to use for a room based on feature flags."""
+    # For now, default to whereby since we don't have feature flags yet
+    return VideoPlatform.WHEREBY

server/reflector/video_platforms/jitsi/__init__.py

@@ -0,0 +1,4 @@
+from .client import JitsiClient, JitsiMeetingData
+from .router import router
+
+__all__ = ["JitsiClient", "JitsiMeetingData", "router"]

server/reflector/video_platforms/jitsi/client.py

@@ -0,0 +1,111 @@
+import hmac
+from datetime import datetime, timezone
+from hashlib import sha256
+from typing import Any, Dict, Optional
+
+import jwt
+
+from reflector.db.rooms import Room, VideoPlatform
+from reflector.settings import settings
+from reflector.utils import generate_uuid4
+
+from ..base import MeetingData, VideoPlatformClient
+
+
+class JitsiMeetingData(MeetingData):
+    @property
+    def user_jwt(self) -> str:
+        return self.extra_data.get("user_jwt", "")
+
+    @property
+    def host_jwt(self) -> str:
+        return self.extra_data.get("host_jwt", "")
+
+    @property
+    def domain(self) -> str:
+        return self.extra_data.get("domain", "")
+
+
+class JitsiClient(VideoPlatformClient):
+    PLATFORM_NAME = VideoPlatform.JITSI
+
+    def _generate_jwt(self, room: str, moderator: bool, exp: datetime) -> str:
+        if not settings.JITSI_JWT_SECRET:
+            raise ValueError("JITSI_JWT_SECRET is required for JWT generation")
+
+        payload = {
+            "aud": settings.JITSI_JWT_AUDIENCE,
+            "iss": settings.JITSI_JWT_ISSUER,
+            "sub": settings.JITSI_DOMAIN,
+            "room": room,
+            "exp": int(exp.timestamp()),
+            "context": {
+                "user": {
+                    "name": "Reflector User",
+                    "moderator": moderator,
+                },
+                "features": {
+                    "recording": True,
+                    "livestreaming": False,
+                },
+            },
+        }
+
+        return jwt.encode(payload, settings.JITSI_JWT_SECRET, algorithm="HS256")
+
+    async def create_meeting(
+        self, room_name_prefix: str, end_date: datetime, room: Room
+    ) -> JitsiMeetingData:
+        jitsi_room = f"reflector-{room.name}-{generate_uuid4()}"
+
+        user_jwt = self._generate_jwt(room=jitsi_room, moderator=False, exp=end_date)
+        host_jwt = self._generate_jwt(room=jitsi_room, moderator=True, exp=end_date)
+
+        room_url = f"https://{settings.JITSI_DOMAIN}/{jitsi_room}?jwt={user_jwt}"
+        host_room_url = f"https://{settings.JITSI_DOMAIN}/{jitsi_room}?jwt={host_jwt}"
+
+        return JitsiMeetingData(
+            meeting_id=generate_uuid4(),
+            room_name=jitsi_room,
+            room_url=room_url,
+            host_room_url=host_room_url,
+            platform=self.PLATFORM_NAME,
+            extra_data={
+                "user_jwt": user_jwt,
+                "host_jwt": host_jwt,
+                "domain": settings.JITSI_DOMAIN,
+            },
+        )
+
+    async def get_room_sessions(self, room_name: str) -> Dict[str, Any]:
+        return {
+            "roomName": room_name,
+            "sessions": [
+                {
+                    "sessionId": generate_uuid4(),
+                    "startTime": datetime.now(tz=timezone.utc).isoformat(),
+                    "participants": [],
+                    "isActive": True,
+                }
+            ],
+        }
+
+    async def delete_room(self, room_name: str) -> bool:
+        return True
+
+    async def upload_logo(self, room_name: str, logo_path: str) -> bool:
+        return True
+
+    def verify_webhook_signature(
+        self, body: bytes, signature: str, timestamp: Optional[str] = None
+    ) -> bool:
+        if not signature or not self.config.webhook_secret:
+            return False
+
+        try:
+            expected = hmac.new(
+                self.config.webhook_secret.encode(), body, sha256
+            ).hexdigest()
+            return hmac.compare_digest(expected, signature)
+        except Exception:
+            return False

server/reflector/video_platforms/jitsi/router.py

@@ -0,0 +1,165 @@
+import hmac
+from datetime import datetime
+from hashlib import sha256
+from typing import Any, Dict
+
+from fastapi import APIRouter, HTTPException, Request
+from pydantic import BaseModel
+
+from reflector.db.meetings import meetings_controller
+from reflector.settings import settings
+
+try:
+    from reflector.video_platforms import create_platform_client
+except ImportError:
+    # PyJWT not yet installed, will be added in final task
+    def create_platform_client(platform: str):
+        return None
+
+
+router = APIRouter()
+
+
+class JitsiWebhookEvent(BaseModel):
+    event: str
+    room: str
+    timestamp: datetime
+    data: Dict[str, Any] = {}
+
+
+class JibriRecordingEvent(BaseModel):
+    room_name: str
+    recording_file: str
+    recording_status: str
+    timestamp: datetime
+
+
+def verify_jitsi_webhook_signature(body: bytes, signature: str) -> bool:
+    """Verify Jitsi webhook signature using HMAC-SHA256."""
+    if not signature or not settings.JITSI_WEBHOOK_SECRET:
+        return False
+
+    try:
+        client = create_platform_client("jitsi")
+        if client is None:
+            # Fallback verification when platform client not available
+            expected = hmac.new(
+                settings.JITSI_WEBHOOK_SECRET.encode(), body, sha256
+            ).hexdigest()
+            return hmac.compare_digest(expected, signature)
+        return client.verify_webhook_signature(body, signature)
+    except Exception:
+        return False
+
+
+@router.post("/jitsi/events")
+async def jitsi_events_webhook(event: JitsiWebhookEvent, request: Request):
+    """
+    Handle Prosody event-sync webhooks from Jitsi Meet.
+
+    Expected event types:
+    - muc-occupant-joined: participant joined the room
+    - muc-occupant-left: participant left the room
+    - jibri-recording-on: recording started
+    - jibri-recording-off: recording stopped
+    """
+    # Verify webhook signature
+    body = await request.body()
+    signature = request.headers.get("x-jitsi-signature", "")
+
+    if not verify_jitsi_webhook_signature(body, signature):
+        raise HTTPException(status_code=401, detail="Invalid webhook signature")
+
+    # Find meeting by room name
+    meeting = await meetings_controller.get_by_room_name(event.room)
+    if not meeting:
+        raise HTTPException(status_code=404, detail="Meeting not found")
+
+    # Handle participant events
+    if event.event == "muc-occupant-joined":
+        # Store event and update participant count
+        await meetings_controller.participant_joined(
+            meeting.id, {"timestamp": event.timestamp, "data": event.data}
+        )
+        current_count = getattr(meeting, "num_clients", 0)
+        await meetings_controller.update_meeting(
+            meeting.id, num_clients=current_count + 1
+        )
+    elif event.event == "muc-occupant-left":
+        # Store event and update participant count
+        await meetings_controller.participant_left(
+            meeting.id, {"timestamp": event.timestamp, "data": event.data}
+        )
+        current_count = getattr(meeting, "num_clients", 0)
+        await meetings_controller.update_meeting(
+            meeting.id, num_clients=max(0, current_count - 1)
+        )
+    elif event.event == "jibri-recording-on":
+        # Store recording started event
+        await meetings_controller.recording_started(
+            meeting.id, {"timestamp": event.timestamp, "data": event.data}
+        )
+    elif event.event == "jibri-recording-off":
+        # Store recording stopped event
+        await meetings_controller.recording_stopped(
+            meeting.id, {"timestamp": event.timestamp, "data": event.data}
+        )
+
+    return {"status": "ok", "event": event.event, "room": event.room}
+
+
+@router.post("/jibri/recording-complete")
+async def jibri_recording_complete(event: JibriRecordingEvent, request: Request):
+    """
+    Handle Jibri recording completion webhook.
+
+    This endpoint is called by the Jibri finalize script when a recording
+    is completed and uploaded to storage.
+    """
+    # Verify webhook signature
+    body = await request.body()
+    signature = request.headers.get("x-jitsi-signature", "")
+
+    if not verify_jitsi_webhook_signature(body, signature):
+        raise HTTPException(status_code=401, detail="Invalid webhook signature")
+
+    # Find meeting by room name
+    meeting = await meetings_controller.get_by_room_name(event.room_name)
+    if not meeting:
+        raise HTTPException(status_code=404, detail="Meeting not found")
+
+    # Store recording completion event
+    await meetings_controller.add_event(
+        meeting.id,
+        "recording_completed",
+        {
+            "recording_file": event.recording_file,
+            "recording_status": event.recording_status,
+            "timestamp": event.timestamp,
+        },
+    )
+
+    # TODO: Trigger recording processing pipeline
+    # This is where we would:
+    # 1. Download the recording file from Jibri storage
+    # 2. Create a transcript record in the database
+    # 3. Queue the audio processing tasks (chunking, transcription, etc.)
+    # 4. Update meeting status to indicate recording is being processed
+
+    return {
+        "status": "ok",
+        "room_name": event.room_name,
+        "recording_file": event.recording_file,
+        "message": "Recording processing queued",
+    }
+
+
+@router.get("/jitsi/health")
+async def jitsi_health_check():
+    """Simple health check endpoint for Jitsi webhook configuration."""
+    return {
+        "status": "ok",
+        "service": "jitsi-webhooks",
+        "timestamp": datetime.now(tz=timezone.utc).isoformat(),
+        "webhook_secret_configured": bool(settings.JITSI_WEBHOOK_SECRET),
+    }

server/reflector/video_platforms/jitsi/tasks.py

@@ -0,0 +1,3 @@
+"""Jitsi-specific worker tasks."""
+
+# Placeholder for Jitsi recording tasks

server/reflector/video_platforms/registry.py

@@ -0,0 +1,56 @@
+from typing import TYPE_CHECKING, Dict, Literal, Type, overload
+
+from .base import VideoPlatformClient, VideoPlatformConfig
+
+if TYPE_CHECKING:
+    from .jitsi import JitsiClient
+    from .whereby import WherebyClient
+
+# Registry of available video platforms
+_PLATFORMS: Dict[str, Type[VideoPlatformClient]] = {}
+
+
+def register_platform(name: str, client_class: Type[VideoPlatformClient]):
+    """Register a video platform implementation."""
+    _PLATFORMS[name.lower()] = client_class
+
+
+@overload
+def get_platform_client(
+    platform: Literal["jitsi"], config: VideoPlatformConfig
+) -> "JitsiClient": ...
+
+
+@overload
+def get_platform_client(
+    platform: Literal["whereby"], config: VideoPlatformConfig
+) -> "WherebyClient": ...
+
+
+def get_platform_client(
+    platform: str, config: VideoPlatformConfig
+) -> VideoPlatformClient:
+    """Get a video platform client instance."""
+    platform_lower = platform.lower()
+    if platform_lower not in _PLATFORMS:
+        raise ValueError(f"Unknown video platform: {platform}")
+
+    client_class = _PLATFORMS[platform_lower]
+    return client_class(config)
+
+
+def get_available_platforms() -> list[str]:
+    """Get list of available platform names."""
+    return list(_PLATFORMS.keys())
+
+
+# Auto-register built-in platforms
+def _register_builtin_platforms():
+    from .jitsi import JitsiClient
+    from .whereby import WherebyClient
+
+    register_platform("jitsi", JitsiClient)
+    register_platform("whereby", WherebyClient)
+
+
+_register_builtin_platforms()

server/reflector/video_platforms/whereby/__init__.py

@@ -0,0 +1,6 @@
+"""Whereby video platform integration."""
+
+from .client import WherebyClient
+from .router import router
+
+__all__ = ["WherebyClient", "router"]

server/reflector/video_platforms/whereby/client.py

@@ -0,0 +1,113 @@
+import hmac
+from datetime import datetime
+from hashlib import sha256
+from typing import Any, Dict, Optional
+
+import httpx
+
+from reflector.db.rooms import Room, VideoPlatform
+from reflector.settings import settings
+
+from ..base import MeetingData, VideoPlatformClient
+
+
+class WherebyClient(VideoPlatformClient):
+    PLATFORM_NAME = VideoPlatform.WHEREBY
+
+    def __init__(self, config):
+        super().__init__(config)
+        self.headers = {
+            "Content-Type": "application/json; charset=utf-8",
+            "Authorization": f"Bearer {self.config.api_key}",
+        }
+        self.timeout = 10
+
+    async def create_meeting(
+        self, room_name_prefix: str, end_date: datetime, room: Room
+    ) -> MeetingData:
+        data = {
+            "isLocked": room.is_locked,
+            "roomNamePrefix": room_name_prefix,
+            "roomNamePattern": "uuid",
+            "roomMode": room.room_mode,
+            "endDate": end_date.isoformat(),
+            "recording": {
+                "type": room.recording_type,
+                "destination": {
+                    "provider": "s3",
+                    "bucket": settings.RECORDING_STORAGE_AWS_BUCKET_NAME,
+                    "accessKeyId": self.config.aws_access_key_id,
+                    "accessKeySecret": self.config.aws_access_key_secret,
+                    "fileFormat": "mp4",
+                },
+                "startTrigger": room.recording_trigger,
+            },
+            "fields": ["hostRoomUrl"],
+        }
+
+        async with httpx.AsyncClient() as client:
+            response = await client.post(
+                f"{self.config.api_url}/meetings",
+                headers=self.headers,
+                json=data,
+                timeout=self.timeout,
+            )
+            response.raise_for_status()
+            meeting_data = response.json()
+
+        return MeetingData(
+            meeting_id=meeting_data["meetingId"],
+            room_name=meeting_data["roomName"],
+            room_url=meeting_data["roomUrl"],
+            host_room_url=meeting_data["hostRoomUrl"],
+            platform=self.PLATFORM_NAME,
+            extra_data={
+                "startDate": meeting_data["startDate"],
+                "endDate": meeting_data["endDate"],
+                "recording": meeting_data.get("recording", {}),
+            },
+        )
+
+    async def get_room_sessions(self, room_name: str) -> Dict[str, Any]:
+        async with httpx.AsyncClient() as client:
+            response = await client.get(
+                f"{self.config.api_url}/insights/room-sessions?roomName={room_name}",
+                headers=self.headers,
+                timeout=self.timeout,
+            )
+            response.raise_for_status()
+            return response.json()
+
+    async def delete_room(self, room_name: str) -> bool:
+        return True
+
+    async def upload_logo(self, room_name: str, logo_path: str) -> bool:
+        try:
+            async with httpx.AsyncClient() as client:
+                with open(logo_path, "rb") as f:
+                    response = await client.put(
+                        f"{self.config.api_url}/rooms{room_name}/theme/logo",
+                        headers={
+                            "Authorization": f"Bearer {self.config.api_key}",
+                        },
+                        timeout=self.timeout,
+                        files={"image": f},
+                    )
+                    response.raise_for_status()
+            return True
+        except Exception:
+            return False
+
+    def verify_webhook_signature(
+        self, body: bytes, signature: str, timestamp: Optional[str] = None
+    ) -> bool:
+        if not signature or not self.config.webhook_secret:
+            return False
+
+        try:
+            expected = hmac.new(
+                self.config.webhook_secret.encode(), body, sha256
+            ).hexdigest()
+            return hmac.compare_digest(expected, signature)
+        except Exception:
+            return False

server/reflector/video_platforms/whereby/tasks.py

@@ -0,0 +1,4 @@
+"""Whereby-specific worker tasks."""
+
+# Placeholder for Whereby-specific background tasks
+# This can be extended with Whereby-specific processing tasks in the future

server/reflector/views/jibri_webhook.py

@@ -0,0 +1,126 @@
+from pathlib import Path
+from typing import Annotated, Any, Dict, Optional
+
+import structlog
+from fastapi import APIRouter, Depends, HTTPException
+from pydantic import BaseModel
+
+import reflector.auth as auth
+from reflector.db.transcripts import SourceKind, transcripts_controller
+from reflector.jibri_events import JitsiEventParser
+from reflector.pipelines.main_file_pipeline import task_pipeline_file_process
+from reflector.settings import settings
+
+logger = structlog.get_logger(__name__)
+
+router = APIRouter(prefix="/api/v1/jibri", tags=["jibri"])
+
+
+class RecordingReadyRequest(BaseModel):
+    session_id: str
+    path: str  # Relative path from recordings directory
+    meeting_url: str
+
+
+@router.post("/recording-ready")
+async def handle_recording_ready(
+    request: RecordingReadyRequest,
+    user: Annotated[Optional[auth.UserInfo], Depends(auth.current_user_optional)],
+) -> Dict[str, Any]:
+    user_id = user["sub"] if user else None
+
+    recordings_base = Path(settings.JIBRI_RECORDINGS_PATH or "/recordings")
+    recording_path = recordings_base / request.path
+
+    if not recording_path.exists():
+        raise HTTPException(
+            status_code=404, detail=f"Recording path not found: {request.path}"
+        )
+
+    recording_file = recording_path / "recording.mp4"
+    events_file = recording_path / "events.jsonl"
+
+    if not recording_file.exists():
+        raise HTTPException(status_code=404, detail="Recording file not found")
+
+    # Parse events if available
+    metadata = {}
+    participant_count = 0
+
+    if events_file.exists():
+        parser = JitsiEventParser()
+        metadata = parser.parse_events_file(str(recording_path))
+        participant_count = len(metadata.get("participants", []))
+        logger.info(
+            "Parsed Jibri events",
+            session_id=request.session_id,
+            event_count=metadata.get("event_count", 0),
+            participant_count=participant_count,
+        )
+    else:
+        logger.warning("No events file found", session_id=request.session_id)
+        metadata = {
+            "room": {"meeting_url": request.meeting_url, "name": request.session_id},
+            "participants": [],
+            "speaker_stats": {},
+            "event_count": 0,
+        }
+
+    # Create transcript using controller
+    title = f"Meeting: {metadata.get('room', {}).get('name', request.session_id)}"
+    transcript = await transcripts_controller.add(
+        name=title,
+        source_kind=SourceKind.FILE,
+        source_language="en",
+        target_language="en",
+        user_id=user_id,
+    )
+
+    # Store Jitsi data in appropriate fields
+    update_data = {}
+
+    # Store participants if available
+    if metadata.get("participants"):
+        update_data["participants"] = metadata["participants"]
+
+    # Store events data (room info, speaker stats, etc.)
+    update_data["events"] = {
+        "jitsi_metadata": metadata,
+        "session_id": request.session_id,
+        "recording_path": str(recording_path),
+        "meeting_url": request.meeting_url,
+    }
+
+    if update_data:
+        await transcripts_controller.update(transcript, update_data)
+
+    # Copy recording file to transcript data path
+    # The pipeline expects the file to be in the transcript's data path
+    upload_file = transcript.data_path / "upload.mp4"
+    upload_file.parent.mkdir(parents=True, exist_ok=True)
+
+    # Create symlink or copy the file
+    import shutil
+
+    shutil.copy2(recording_file, upload_file)
+
+    # Update status to uploaded
+    await transcripts_controller.update(transcript, {"status": "uploaded"})
+
+    # Trigger processing pipeline
+    task_pipeline_file_process.delay(transcript_id=transcript.id)
+
+    logger.info(
+        "Jibri recording ready for processing",
+        transcript_id=transcript.id,
+        session_id=request.session_id,
+        participant_count=participant_count,
+    )
+
+    return {
+        "status": "accepted",
+        "transcript_id": transcript.id,
+        "session_id": request.session_id,
+        "events_found": events_file.exists(),
+        "participant_count": participant_count,
+    }

server/reflector/views/rooms.py

@@ -12,9 +12,12 @@ from pydantic import BaseModel
 import reflector.auth as auth
 from reflector.db import get_database
 from reflector.db.meetings import meetings_controller
-from reflector.db.rooms import rooms_controller
+from reflector.db.rooms import VideoPlatform, rooms_controller
 from reflector.settings import settings
-from reflector.whereby import create_meeting, upload_logo
+from reflector.video_platforms.factory import (
+    create_platform_client,
+)
+from reflector.worker.webhook import test_webhook
 
 logger = logging.getLogger(__name__)
 
@@ -22,7 +25,6 @@ router = APIRouter()
 
 
 def parse_datetime_with_timezone(iso_string: str) -> datetime:
-    """Parse ISO datetime string and ensure timezone awareness (defaults to UTC if naive)."""
     dt = datetime.fromisoformat(iso_string)
     if dt.tzinfo is None:
         dt = dt.replace(tzinfo=timezone.utc)
@@ -42,6 +44,12 @@ class Room(BaseModel):
     recording_type: str
     recording_trigger: str
     is_shared: bool
+    platform: VideoPlatform = VideoPlatform.WHEREBY
+
+
+class RoomDetails(Room):
+    webhook_url: str | None
+    webhook_secret: str | None
 
 
 class Meeting(BaseModel):
@@ -64,6 +72,9 @@ class CreateRoom(BaseModel):
     recording_type: str
     recording_trigger: str
     is_shared: bool
+    webhook_url: str
+    webhook_secret: str
+    platform: VideoPlatform
 
 
 class UpdateRoom(BaseModel):
@@ -76,16 +87,27 @@ class UpdateRoom(BaseModel):
     recording_type: str
     recording_trigger: str
     is_shared: bool
+    webhook_url: str
+    webhook_secret: str
+    platform: VideoPlatform
 
 
 class DeletionStatus(BaseModel):
     status: str
 
 
-@router.get("/rooms", response_model=Page[Room])
+class WebhookTestResult(BaseModel):
+    success: bool
+    message: str = ""
+    error: str = ""
+    status_code: int | None = None
+    response_preview: str | None = None
+
+
+@router.get("/rooms", response_model=Page[RoomDetails])
 async def rooms_list(
     user: Annotated[Optional[auth.UserInfo], Depends(auth.current_user_optional)],
-) -> list[Room]:
+) -> list[RoomDetails]:
     if not user and not settings.PUBLIC_MODE:
         raise HTTPException(status_code=401, detail="Not authenticated")
 
@@ -99,6 +121,18 @@ async def rooms_list(
     )
 
 
+@router.get("/rooms/{room_id}", response_model=RoomDetails)
+async def rooms_get(
+    room_id: str,
+    user: Annotated[Optional[auth.UserInfo], Depends(auth.current_user_optional)],
+):
+    user_id = user["sub"] if user else None
+    room = await rooms_controller.get_by_id_for_http(room_id, user_id=user_id)
+    if not room:
+        raise HTTPException(status_code=404, detail="Room not found")
+    return room
+
+
 @router.post("/rooms", response_model=Room)
 async def rooms_create(
     room: CreateRoom,
@@ -117,10 +151,13 @@ async def rooms_create(
         recording_type=room.recording_type,
         recording_trigger=room.recording_trigger,
         is_shared=room.is_shared,
+        webhook_url=room.webhook_url,
+        webhook_secret=room.webhook_secret,
+        platform=room.platform,
     )
 
 
-@router.patch("/rooms/{room_id}", response_model=Room)
+@router.patch("/rooms/{room_id}", response_model=RoomDetails)
 async def rooms_update(
     room_id: str,
     info: UpdateRoom,
@@ -164,42 +201,49 @@ async def rooms_create_meeting(
     if meeting is None:
         end_date = current_time + timedelta(hours=8)
 
-        whereby_meeting = await create_meeting("", end_date=end_date, room=room)
-        await upload_logo(whereby_meeting["roomName"], "./images/logo.png")
+        platform = room.platform
+        client = create_platform_client(platform)
 
-        # Now try to save to database
+        platform_meeting = await client.create_meeting("", end_date=end_date, room=room)
+        await client.upload_logo(platform_meeting.room_name, "./images/logo.png")
+
+        meeting_data = {
+            "meeting_id": platform_meeting.meeting_id,
+            "room_name": platform_meeting.room_name,
+            "room_url": platform_meeting.room_url,
+            "host_room_url": platform_meeting.host_room_url,
+            "start_date": current_time,
+            "end_date": end_date,
+        }
         try:
             meeting = await meetings_controller.create(
-                id=whereby_meeting["meetingId"],
-                room_name=whereby_meeting["roomName"],
-                room_url=whereby_meeting["roomUrl"],
-                host_room_url=whereby_meeting["hostRoomUrl"],
-                start_date=parse_datetime_with_timezone(whereby_meeting["startDate"]),
-                end_date=parse_datetime_with_timezone(whereby_meeting["endDate"]),
+                id=meeting_data["meeting_id"],
+                room_name=meeting_data["room_name"],
+                room_url=meeting_data["room_url"],
+                host_room_url=meeting_data["host_room_url"],
+                start_date=meeting_data["start_date"],
+                end_date=meeting_data["end_date"],
                 user_id=user_id,
                 room=room,
             )
         except (asyncpg.exceptions.UniqueViolationError, sqlite3.IntegrityError):
-            # Another request already created a meeting for this room
-            # Log this race condition occurrence
             logger.info(
                 "Race condition detected for room %s - fetching existing meeting",
                 room.name,
             )
             logger.warning(
-                "Whereby meeting %s was created but not used (resource leak) for room %s",
-                whereby_meeting["meetingId"],
+                "Platform meeting %s was created but not used (resource leak) for room %s",
+                meeting_data["meeting_id"],
                 room.name,
             )
-
-            # Fetch the meeting that was created by the other request
             meeting = await meetings_controller.get_active(
                 room=room, current_time=current_time
             )
             if meeting is None:
-                # Edge case: meeting was created but expired/deleted between checks
                 logger.error(
-                    "Meeting disappeared after race condition for room %s", room.name
+                    "Meeting disappeared after race condition for room %s",
+                    room.name,
+                    exc_info=True,
                 )
                 raise HTTPException(
                     status_code=503, detail="Unable to join meeting - please try again"
@@ -209,3 +253,23 @@ async def rooms_create_meeting(
         meeting.host_room_url = ""
 
     return meeting
+
+
+@router.post("/rooms/{room_id}/webhook/test", response_model=WebhookTestResult)
+async def rooms_test_webhook(
+    room_id: str,
+    user: Annotated[Optional[auth.UserInfo], Depends(auth.current_user_optional)],
+):
+    user_id = user["sub"] if user else None
+
+    room = await rooms_controller.get_by_id(room_id)
+    if not room:
+        raise HTTPException(status_code=404, detail="Room not found")
+
+    if user_id and room.user_id != user_id:
+        raise HTTPException(
+            status_code=403, detail="Not authorized to test this room's webhook"
+        )
+
+    result = await test_webhook(room_id)
+    return WebhookTestResult(**result)

server/reflector/views/transcripts.py

@@ -5,7 +5,7 @@ from fastapi import APIRouter, Depends, HTTPException, Query
 from fastapi_pagination import Page
 from fastapi_pagination.ext.databases import apaginate
 from jose import jwt
-from pydantic import BaseModel, Field, field_serializer
+from pydantic import BaseModel, Field, constr, field_serializer
 
 import reflector.auth as auth
 from reflector.db import get_database
@@ -19,14 +19,15 @@ from reflector.db.search import (
     SearchOffsetBase,
     SearchParameters,
     SearchQuery,
-    SearchQueryBase,
     SearchResult,
     SearchTotal,
     search_controller,
+    search_query_adapter,
 )
 from reflector.db.transcripts import (
     SourceKind,
     TranscriptParticipant,
+    TranscriptStatus,
     TranscriptTopic,
     transcripts_controller,
 )
@@ -63,7 +64,7 @@ class GetTranscriptMinimal(BaseModel):
     id: str
     user_id: str | None
     name: str
-    status: str
+    status: TranscriptStatus
     locked: bool
     duration: float
     title: str | None
@@ -96,6 +97,7 @@ class CreateTranscript(BaseModel):
     name: str
     source_language: str = Field("en")
     target_language: str = Field("en")
+    source_kind: SourceKind | None = None
 
 
 class UpdateTranscript(BaseModel):
@@ -114,7 +116,19 @@ class DeletionStatus(BaseModel):
     status: str
 
 
-SearchQueryParam = Annotated[SearchQueryBase, Query(description="Search query text")]
+SearchQueryParamBase = constr(min_length=0, strip_whitespace=True)
+SearchQueryParam = Annotated[
+    SearchQueryParamBase, Query(description="Search query text")
+]
+
+
+# http and api standards accept "q="; we would like to handle it as the absence of query, not as "empty string query"
+def parse_search_query_param(q: SearchQueryParam) -> SearchQuery | None:
+    if q == "":
+        return None
+    return search_query_adapter.validate_python(q)
+
+
 SearchLimitParam = Annotated[SearchLimitBase, Query(description="Results per page")]
 SearchOffsetParam = Annotated[
     SearchOffsetBase, Query(description="Number of results to skip")
@@ -124,7 +138,7 @@ SearchOffsetParam = Annotated[
 class SearchResponse(BaseModel):
     results: list[SearchResult]
     total: SearchTotal
-    query: SearchQuery
+    query: SearchQuery | None = None
     limit: SearchLimit
     offset: SearchOffset
 
@@ -174,7 +188,7 @@ async def transcripts_search(
     user_id = user["sub"] if user else None
 
     search_params = SearchParameters(
-        query_text=q,
+        query_text=parse_search_query_param(q),
         limit=limit,
         offset=offset,
         user_id=user_id,
@@ -201,7 +215,7 @@ async def transcripts_create(
     user_id = user["sub"] if user else None
     return await transcripts_controller.add(
         info.name,
-        source_kind=SourceKind.LIVE,
+        source_kind=info.source_kind or SourceKind.LIVE,
         source_language=info.source_language,
         target_language=info.target_language,
         user_id=user_id,
@@ -336,8 +350,6 @@ async def transcript_update(
     transcript = await transcripts_controller.get_by_id_for_http(
         transcript_id, user_id=user_id
     )
-    if not transcript:
-        raise HTTPException(status_code=404, detail="Transcript not found")
     values = info.dict(exclude_unset=True)
     updated_transcript = await transcripts_controller.update(transcript, values)
     return updated_transcript

server/reflector/views/transcripts_process.py

@@ -6,7 +6,7 @@ from pydantic import BaseModel
 
 import reflector.auth as auth
 from reflector.db.transcripts import transcripts_controller
-from reflector.pipelines.main_live_pipeline import task_pipeline_process
+from reflector.pipelines.main_file_pipeline import task_pipeline_file_process
 
 router = APIRouter()
 
@@ -34,13 +34,13 @@ async def transcript_process(
         )
 
     if task_is_scheduled_or_active(
-        "reflector.pipelines.main_live_pipeline.task_pipeline_process",
+        "reflector.pipelines.main_file_pipeline.task_pipeline_file_process",
         transcript_id=transcript_id,
     ):
         return ProcessStatus(status="already running")
 
     # schedule a background task process the file
-    task_pipeline_process.delay(transcript_id=transcript_id)
+    task_pipeline_file_process.delay(transcript_id=transcript_id)
 
     return ProcessStatus(status="ok")
 

server/reflector/views/transcripts_upload.py

@@ -6,7 +6,7 @@ from pydantic import BaseModel
 
 import reflector.auth as auth
 from reflector.db.transcripts import transcripts_controller
-from reflector.pipelines.main_live_pipeline import task_pipeline_process
+from reflector.pipelines.main_file_pipeline import task_pipeline_file_process
 
 router = APIRouter()
 
@@ -92,6 +92,6 @@ async def transcript_record_upload(
     await transcripts_controller.update(transcript, {"status": "uploaded"})
 
     # launch a background task to process the file
-    task_pipeline_process.delay(transcript_id=transcript_id)
+    task_pipeline_file_process.delay(transcript_id=transcript_id)
 
     return UploadStatus(status="ok")

server/reflector/whereby.py

@@ -1,18 +1,60 @@
+import logging
 from datetime import datetime
 
 import httpx
 
 from reflector.db.rooms import Room
 from reflector.settings import settings
+from reflector.utils.string import parse_non_empty_string
+
+logger = logging.getLogger(__name__)
+
+
+def _get_headers():
+    api_key = parse_non_empty_string(
+        settings.WHEREBY_API_KEY, "WHEREBY_API_KEY value is required."
+    )
+    return {
+        "Content-Type": "application/json; charset=utf-8",
+        "Authorization": f"Bearer {api_key}",
+    }
+
 
-HEADERS = {
-    "Content-Type": "application/json; charset=utf-8",
-    "Authorization": f"Bearer {settings.WHEREBY_API_KEY}",
-}
 TIMEOUT = 10  # seconds
 
 
+def _get_whereby_s3_auth():
+    errors = []
+    try:
+        bucket_name = parse_non_empty_string(
+            settings.RECORDING_STORAGE_AWS_BUCKET_NAME,
+            "RECORDING_STORAGE_AWS_BUCKET_NAME value is required.",
+        )
+    except Exception as e:
+        errors.append(e)
+    try:
+        key_id = parse_non_empty_string(
+            settings.AWS_WHEREBY_ACCESS_KEY_ID,
+            "AWS_WHEREBY_ACCESS_KEY_ID value is required.",
+        )
+    except Exception as e:
+        errors.append(e)
+    try:
+        key_secret = parse_non_empty_string(
+            settings.AWS_WHEREBY_ACCESS_KEY_SECRET,
+            "AWS_WHEREBY_ACCESS_KEY_SECRET value is required.",
+        )
+    except Exception as e:
+        errors.append(e)
+    if len(errors) > 0:
+        raise Exception(
+            f"Failed to get Whereby auth settings: {', '.join(str(e) for e in errors)}"
+        )
+    return bucket_name, key_id, key_secret
+
+
 async def create_meeting(room_name_prefix: str, end_date: datetime, room: Room):
+    s3_bucket_name, s3_key_id, s3_key_secret = _get_whereby_s3_auth()
     data = {
         "isLocked": room.is_locked,
         "roomNamePrefix": room_name_prefix,
@@ -23,23 +65,26 @@ async def create_meeting(room_name_prefix: str, end_date: datetime, room: Room):
             "type": room.recording_type,
             "destination": {
                 "provider": "s3",
-                "bucket": settings.RECORDING_STORAGE_AWS_BUCKET_NAME,
-                "accessKeyId": settings.AWS_WHEREBY_ACCESS_KEY_ID,
-                "accessKeySecret": settings.AWS_WHEREBY_ACCESS_KEY_SECRET,
+                "bucket": s3_bucket_name,
+                "accessKeyId": s3_key_id,
+                "accessKeySecret": s3_key_secret,
                 "fileFormat": "mp4",
             },
             "startTrigger": room.recording_trigger,
         },
         "fields": ["hostRoomUrl"],
     }
-
     async with httpx.AsyncClient() as client:
         response = await client.post(
             f"{settings.WHEREBY_API_URL}/meetings",
-            headers=HEADERS,
+            headers=_get_headers(),
             json=data,
             timeout=TIMEOUT,
         )
+        if response.status_code == 403:
+            logger.warning(
+                f"Failed to create meeting: access denied on Whereby: {response.text}"
+            )
         response.raise_for_status()
         return response.json()
 
@@ -48,7 +93,7 @@ async def get_room_sessions(room_name: str):
     async with httpx.AsyncClient() as client:
         response = await client.get(
             f"{settings.WHEREBY_API_URL}/insights/room-sessions?roomName={room_name}",
-            headers=HEADERS,
+            headers=_get_headers(),
             timeout=TIMEOUT,
         )
         response.raise_for_status()

server/reflector/worker/app.py

@@ -19,6 +19,8 @@ else:
             "reflector.pipelines.main_live_pipeline",
             "reflector.worker.healthcheck",
             "reflector.worker.process",
+            "reflector.worker.cleanup",
+            "reflector.worker.jitsi_events",
         ]
     )
 
@@ -32,12 +34,26 @@ else:
             "task": "reflector.worker.process.process_meetings",
             "schedule": float(settings.SQS_POLLING_TIMEOUT_SECONDS),
         },
+        "process_jitsi_events": {
+            "task": "reflector.worker.jitsi_events.process_jitsi_events",
+            "schedule": 5.0,  # Process every 5 seconds
+        },
         "reprocess_failed_recordings": {
             "task": "reflector.worker.process.reprocess_failed_recordings",
             "schedule": crontab(hour=5, minute=0),  # Midnight EST
         },
     }
 
+    if settings.PUBLIC_MODE:
+        app.conf.beat_schedule["cleanup_old_public_data"] = {
+            "task": "reflector.worker.cleanup.cleanup_old_public_data_task",
+            "schedule": crontab(hour=3, minute=0),
+        }
+        logger.info(
+            "Public mode cleanup enabled",
+            retention_days=settings.PUBLIC_DATA_RETENTION_DAYS,
+        )
+
     if settings.HEALTHCHECK_URL:
         app.conf.beat_schedule["healthcheck_ping"] = {
             "task": "reflector.worker.healthcheck.healthcheck_ping",

server/reflector/worker/cleanup.py

@@ -0,0 +1,156 @@
+"""
+Main task for cleanup old public data.
+
+Deletes old anonymous transcripts and their associated meetings/recordings.
+Transcripts are the main entry point - any associated data is also removed.
+"""
+
+import asyncio
+from datetime import datetime, timedelta, timezone
+from typing import TypedDict
+
+import structlog
+from celery import shared_task
+from databases import Database
+from pydantic.types import PositiveInt
+
+from reflector.asynctask import asynctask
+from reflector.db import get_database
+from reflector.db.meetings import meetings
+from reflector.db.recordings import recordings
+from reflector.db.transcripts import transcripts, transcripts_controller
+from reflector.settings import settings
+from reflector.storage import get_recordings_storage
+
+logger = structlog.get_logger(__name__)
+
+
+class CleanupStats(TypedDict):
+    """Statistics for cleanup operation."""
+
+    transcripts_deleted: int
+    meetings_deleted: int
+    recordings_deleted: int
+    errors: list[str]
+
+
+async def delete_single_transcript(
+    db: Database, transcript_data: dict, stats: CleanupStats
+):
+    transcript_id = transcript_data["id"]
+    meeting_id = transcript_data["meeting_id"]
+    recording_id = transcript_data["recording_id"]
+
+    try:
+        async with db.transaction(isolation="serializable"):
+            if meeting_id:
+                await db.execute(meetings.delete().where(meetings.c.id == meeting_id))
+                stats["meetings_deleted"] += 1
+                logger.info("Deleted associated meeting", meeting_id=meeting_id)
+
+            if recording_id:
+                recording = await db.fetch_one(
+                    recordings.select().where(recordings.c.id == recording_id)
+                )
+                if recording:
+                    try:
+                        await get_recordings_storage().delete_file(
+                            recording["object_key"]
+                        )
+                    except Exception as storage_error:
+                        logger.warning(
+                            "Failed to delete recording from storage",
+                            recording_id=recording_id,
+                            object_key=recording["object_key"],
+                            error=str(storage_error),
+                        )
+
+                    await db.execute(
+                        recordings.delete().where(recordings.c.id == recording_id)
+                    )
+                    stats["recordings_deleted"] += 1
+                    logger.info(
+                        "Deleted associated recording", recording_id=recording_id
+                    )
+
+            await transcripts_controller.remove_by_id(transcript_id)
+            stats["transcripts_deleted"] += 1
+            logger.info(
+                "Deleted transcript",
+                transcript_id=transcript_id,
+                created_at=transcript_data["created_at"].isoformat(),
+            )
+    except Exception as e:
+        error_msg = f"Failed to delete transcript {transcript_id}: {str(e)}"
+        logger.error(error_msg, exc_info=e)
+        stats["errors"].append(error_msg)
+
+
+async def cleanup_old_transcripts(
+    db: Database, cutoff_date: datetime, stats: CleanupStats
+):
+    """Delete old anonymous transcripts and their associated recordings/meetings."""
+    query = transcripts.select().where(
+        (transcripts.c.created_at < cutoff_date) & (transcripts.c.user_id.is_(None))
+    )
+    old_transcripts = await db.fetch_all(query)
+
+    logger.info(f"Found {len(old_transcripts)} old transcripts to delete")
+
+    for transcript_data in old_transcripts:
+        await delete_single_transcript(db, transcript_data, stats)
+
+
+def log_cleanup_results(stats: CleanupStats):
+    logger.info(
+        "Cleanup completed",
+        transcripts_deleted=stats["transcripts_deleted"],
+        meetings_deleted=stats["meetings_deleted"],
+        recordings_deleted=stats["recordings_deleted"],
+        errors_count=len(stats["errors"]),
+    )
+
+    if stats["errors"]:
+        logger.warning(
+            "Cleanup completed with errors",
+            errors=stats["errors"][:10],
+        )
+
+
+async def cleanup_old_public_data(
+    days: PositiveInt | None = None,
+) -> CleanupStats | None:
+    if days is None:
+        days = settings.PUBLIC_DATA_RETENTION_DAYS
+
+    if not settings.PUBLIC_MODE:
+        logger.info("Skipping cleanup - not a public instance")
+        return None
+
+    cutoff_date = datetime.now(timezone.utc) - timedelta(days=days)
+    logger.info(
+        "Starting cleanup of old public data",
+        cutoff_date=cutoff_date.isoformat(),
+    )
+
+    stats: CleanupStats = {
+        "transcripts_deleted": 0,
+        "meetings_deleted": 0,
+        "recordings_deleted": 0,
+        "errors": [],
+    }
+
+    db = get_database()
+    await cleanup_old_transcripts(db, cutoff_date, stats)
+
+    log_cleanup_results(stats)
+    return stats
+
+
+@shared_task(
+    autoretry_for=(Exception,),
+    retry_kwargs={"max_retries": 3, "countdown": 300},
+)
+@asynctask
+def cleanup_old_public_data_task(days: int | None = None):
+    asyncio.run(cleanup_old_public_data(days=days))

server/reflector/worker/jitsi_events.py

@@ -0,0 +1,281 @@
+"""
+Celery tasks for consuming Jitsi events from Redis queues.
+"""
+
+import json
+from datetime import datetime
+from typing import Any, Dict
+
+import redis
+import structlog
+from sqlalchemy.orm import Session
+
+from reflector.database import get_db_sync
+from reflector.models import Meeting, Transcript
+from reflector.settings import settings
+from reflector.worker.app import app
+
+logger = structlog.get_logger(__name__)
+
+
+class JitsiEventProcessor:
+    """Process Jitsi events from Redis queues."""
+
+    def __init__(self):
+        self.redis_client = redis.Redis(
+            host=settings.REDIS_HOST or "redis",
+            port=settings.REDIS_PORT or 6379,
+            decode_responses=True,
+        )
+        self.participants = {}  # room_name -> {jid: participant_info}
+        self.speaker_stats = {}  # room_name -> {jid: stats}
+
+    def process_participant_joined(self, data: Dict[str, Any], db: Session):
+        """Track participant joining a room."""
+        room_name = data["room_name"]
+        participant = {
+            "jid": data["participant_jid"],
+            "nick": data["participant_nick"],
+            "id": data["participant_id"],
+            "is_moderator": data.get("is_moderator", False),
+            "joined_at": datetime.now(),
+        }
+
+        if room_name not in self.participants:
+            self.participants[room_name] = {}
+
+        self.participants[room_name][participant["jid"]] = participant
+
+        logger.info(
+            "Participant joined",
+            room=room_name,
+            participant=participant["nick"],
+            total_participants=len(self.participants[room_name]),
+        )
+
+        # Update meeting in database if exists
+        meeting = (
+            db.query(Meeting)
+            .filter(
+                Meeting.room_name == room_name,
+                Meeting.status.in_(["active", "pending"]),
+            )
+            .first()
+        )
+
+        if meeting:
+            # Store participant info in meeting metadata
+            metadata = meeting.metadata or {}
+            if "participants" not in metadata:
+                metadata["participants"] = []
+
+            metadata["participants"].append(
+                {
+                    "id": participant["id"],
+                    "name": participant["nick"],
+                    "joined_at": participant["joined_at"].isoformat(),
+                    "is_moderator": participant["is_moderator"],
+                }
+            )
+
+            meeting.metadata = metadata
+            db.commit()
+
+    def process_participant_left(self, data: Dict[str, Any], db: Session):
+        """Track participant leaving a room."""
+        room_name = data["room_name"]
+        participant_jid = data["participant_jid"]
+
+        if room_name in self.participants:
+            if participant_jid in self.participants[room_name]:
+                participant = self.participants[room_name][participant_jid]
+                participant["left_at"] = datetime.now()
+
+                logger.info(
+                    "Participant left",
+                    room=room_name,
+                    participant=participant["nick"],
+                    duration=(
+                        participant["left_at"] - participant["joined_at"]
+                    ).total_seconds(),
+                )
+
+                # Update meeting in database
+                meeting = (
+                    db.query(Meeting)
+                    .filter(
+                        Meeting.room_name == room_name,
+                        Meeting.status.in_(["active", "pending"]),
+                    )
+                    .first()
+                )
+
+                if meeting and meeting.metadata and "participants" in meeting.metadata:
+                    for p in meeting.metadata["participants"]:
+                        if p["id"] == participant["id"]:
+                            p["left_at"] = participant["left_at"].isoformat()
+                            break
+                    db.commit()
+
+    def process_speaker_stats(self, data: Dict[str, Any], db: Session):
+        """Update speaker statistics."""
+        room_name = data["room_jid"].split("@")[0]
+        self.speaker_stats[room_name] = data["stats"]
+
+        logger.debug(
+            "Speaker stats updated", room=room_name, speakers=len(data["stats"])
+        )
+
+    def process_recording_completed(self, data: Dict[str, Any], db: Session):
+        """Process completed recording with all metadata."""
+        room_name = data["room_name"]
+        meeting_url = data["meeting_url"]
+        recording_path = data["recording_path"]
+        recording_file = data["recording_file"]
+
+        logger.info(
+            "Recording completed", room=room_name, url=meeting_url, path=recording_path
+        )
+
+        # Get participant data for this room
+        participants = self.participants.get(room_name, {})
+        speaker_stats = self.speaker_stats.get(room_name, {})
+
+        # Create transcript record with full metadata
+        transcript = Transcript(
+            title=f"Recording: {room_name}",
+            source_url=meeting_url,
+            metadata={
+                "jitsi": {
+                    "room_name": room_name,
+                    "meeting_url": meeting_url,
+                    "recording_path": recording_path,
+                    "participants": [
+                        {
+                            "id": p["id"],
+                            "name": p["nick"],
+                            "joined_at": p["joined_at"].isoformat(),
+                            "left_at": p.get("left_at", datetime.now()).isoformat(),
+                            "is_moderator": p["is_moderator"],
+                            "speaking_time": speaker_stats.get(p["jid"], {}).get(
+                                "total_time", 0
+                            ),
+                        }
+                        for p in participants.values()
+                    ],
+                    "speaker_stats": speaker_stats,
+                }
+            },
+            status="pending",
+        )
+        db.add(transcript)
+        db.commit()
+
+        # Trigger processing pipeline
+        from reflector.pipelines.main_transcript_pipeline import TranscriptMainPipeline
+
+        pipeline = TranscriptMainPipeline()
+        pipeline.create(transcript.id, recording_file)
+
+        # Clean up room data
+        self.participants.pop(room_name, None)
+        self.speaker_stats.pop(room_name, None)
+
+        logger.info(
+            "Transcript created",
+            transcript_id=transcript.id,
+            participants=len(participants),
+            has_speaker_stats=bool(speaker_stats),
+        )
+
+
+processor = JitsiEventProcessor()
+
+
+@app.task(name="reflector.worker.jitsi_events.process_jitsi_events")
+def process_jitsi_events():
+    """
+    Process Jitsi events from Redis queue.
+    This should be called periodically by Celery Beat.
+    """
+    db = next(get_db_sync())
+    processed = 0
+
+    try:
+        # Process up to 100 events per run
+        for _ in range(100):
+            # Pop event from queue (blocking with 1 second timeout)
+            event_data = processor.redis_client.brpop(
+                ["jitsi:events:queue", "jitsi:recordings:queue"], timeout=1
+            )
+
+            if not event_data:
+                break
+
+            queue_name, event_json = event_data
+            event = json.loads(event_json)
+
+            event_type = event["type"]
+            data = event["data"]
+
+            logger.debug(f"Processing event: {event_type}")
+
+            # Route to appropriate processor
+            if event_type == "participant_joined":
+                processor.process_participant_joined(data, db)
+            elif event_type == "participant_left":
+                processor.process_participant_left(data, db)
+            elif event_type == "speaker_stats_update":
+                processor.process_speaker_stats(data, db)
+            elif event_type == "recording_completed":
+                processor.process_recording_completed(data, db)
+            else:
+                logger.warning(f"Unknown event type: {event_type}")
+
+            processed += 1
+
+        if processed > 0:
+            logger.info(f"Processed {processed} Jitsi events")
+
+    except Exception as e:
+        logger.error(f"Error processing Jitsi events: {e}")
+        raise
+    finally:
+        db.close()
+
+    return processed
+
+
+@app.task(name="reflector.worker.jitsi_events.consume_jitsi_stream")
+def consume_jitsi_stream():
+    """
+    Alternative: Use Redis Streams for more reliable event processing.
+    Redis Streams provide better guarantees and consumer groups.
+    """
+    db = next(get_db_sync())
+
+    try:
+        # Read from stream with consumer group
+        events = processor.redis_client.xreadgroup(
+            "reflector-consumers",
+            "worker-1",
+            {"jitsi:events": ">"},
+            count=10,
+            block=1000,
+        )
+
+        for stream_name, messages in events:
+            for message_id, data in messages:
+                event = json.loads(data[b"event"])
+                # Process event...
+
+                # Acknowledge message
+                processor.redis_client.xack(
+                    stream_name, "reflector-consumers", message_id
+                )
+
+    except Exception as e:
+        logger.error(f"Error consuming stream: {e}")
+        raise
+    finally:
+        db.close()

server/reflector/worker/process.py

@@ -17,7 +17,7 @@ from reflector.db.transcripts import SourceKind, transcripts_controller
 from reflector.pipelines.main_file_pipeline import task_pipeline_file_process
 from reflector.pipelines.main_live_pipeline import asynctask
 from reflector.settings import settings
-from reflector.whereby import get_room_sessions
+from reflector.video_platforms.factory import create_platform_client
 
 logger = structlog.wrap_logger(get_task_logger(__name__))
 
@@ -155,11 +155,18 @@ async def process_meetings():
         if end_date.tzinfo is None:
             end_date = end_date.replace(tzinfo=timezone.utc)
         if end_date > datetime.now(timezone.utc):
-            response = await get_room_sessions(meeting.room_name)
-            room_sessions = response.get("results", [])
-            is_active = not room_sessions or any(
-                rs["endedAt"] is None for rs in room_sessions
-            )
+            # Get room sessions using platform client
+            platform = getattr(meeting, "platform", "whereby")
+            client = create_platform_client(platform)
+            if client:
+                response = await client.get_room_sessions(meeting.room_name)
+                room_sessions = response.get("results", [])
+                is_active = not room_sessions or any(
+                    rs["endedAt"] is None for rs in room_sessions
+                )
+            else:
+                # Fallback: assume meeting is still active if we can't check
+                is_active = True
         if not is_active:
             await meetings_controller.update_meeting(meeting.id, is_active=False)
             logger.info("Meeting %s is deactivated", meeting.id)

server/reflector/worker/webhook.py

@@ -0,0 +1,258 @@
+"""Webhook task for sending transcript notifications."""
+
+import hashlib
+import hmac
+import json
+import uuid
+from datetime import datetime, timezone
+
+import httpx
+import structlog
+from celery import shared_task
+from celery.utils.log import get_task_logger
+
+from reflector.db.rooms import rooms_controller
+from reflector.db.transcripts import transcripts_controller
+from reflector.pipelines.main_live_pipeline import asynctask
+from reflector.settings import settings
+from reflector.utils.webvtt import topics_to_webvtt
+
+logger = structlog.wrap_logger(get_task_logger(__name__))
+
+
+def generate_webhook_signature(payload: bytes, secret: str, timestamp: str) -> str:
+    """Generate HMAC signature for webhook payload."""
+    signed_payload = f"{timestamp}.{payload.decode('utf-8')}"
+    hmac_obj = hmac.new(
+        secret.encode("utf-8"),
+        signed_payload.encode("utf-8"),
+        hashlib.sha256,
+    )
+    return hmac_obj.hexdigest()
+
+
+@shared_task(
+    bind=True,
+    max_retries=30,
+    default_retry_delay=60,
+    retry_backoff=True,
+    retry_backoff_max=3600,  # Max 1 hour between retries
+)
+@asynctask
+async def send_transcript_webhook(
+    self,
+    transcript_id: str,
+    room_id: str,
+    event_id: str,
+):
+    log = logger.bind(
+        transcript_id=transcript_id,
+        room_id=room_id,
+        retry_count=self.request.retries,
+    )
+
+    try:
+        # Fetch transcript and room
+        transcript = await transcripts_controller.get_by_id(transcript_id)
+        if not transcript:
+            log.error("Transcript not found, skipping webhook")
+            return
+
+        room = await rooms_controller.get_by_id(room_id)
+        if not room:
+            log.error("Room not found, skipping webhook")
+            return
+
+        if not room.webhook_url:
+            log.info("No webhook URL configured for room, skipping")
+            return
+
+        # Generate WebVTT content from topics
+        topics_data = []
+
+        if transcript.topics:
+            # Build topics data with diarized content per topic
+            for topic in transcript.topics:
+                topic_webvtt = topics_to_webvtt([topic]) if topic.words else ""
+                topics_data.append(
+                    {
+                        "title": topic.title,
+                        "summary": topic.summary,
+                        "timestamp": topic.timestamp,
+                        "duration": topic.duration,
+                        "webvtt": topic_webvtt,
+                    }
+                )
+
+        # Build webhook payload
+        frontend_url = f"{settings.UI_BASE_URL}/transcripts/{transcript.id}"
+        participants = [
+            {"id": p.id, "name": p.name, "speaker": p.speaker}
+            for p in (transcript.participants or [])
+        ]
+        payload_data = {
+            "event": "transcript.completed",
+            "event_id": event_id,
+            "timestamp": datetime.now(timezone.utc).isoformat(),
+            "transcript": {
+                "id": transcript.id,
+                "room_id": transcript.room_id,
+                "created_at": transcript.created_at.isoformat(),
+                "duration": transcript.duration,
+                "title": transcript.title,
+                "short_summary": transcript.short_summary,
+                "long_summary": transcript.long_summary,
+                "webvtt": transcript.webvtt,
+                "topics": topics_data,
+                "participants": participants,
+                "source_language": transcript.source_language,
+                "target_language": transcript.target_language,
+                "status": transcript.status,
+                "frontend_url": frontend_url,
+            },
+            "room": {
+                "id": room.id,
+                "name": room.name,
+            },
+        }
+
+        # Convert to JSON
+        payload_json = json.dumps(payload_data, separators=(",", ":"))
+        payload_bytes = payload_json.encode("utf-8")
+
+        # Generate signature if secret is configured
+        headers = {
+            "Content-Type": "application/json",
+            "User-Agent": "Reflector-Webhook/1.0",
+            "X-Webhook-Event": "transcript.completed",
+            "X-Webhook-Retry": str(self.request.retries),
+        }
+
+        if room.webhook_secret:
+            timestamp = str(int(datetime.now(timezone.utc).timestamp()))
+            signature = generate_webhook_signature(
+                payload_bytes, room.webhook_secret, timestamp
+            )
+            headers["X-Webhook-Signature"] = f"t={timestamp},v1={signature}"
+
+        # Send webhook with timeout
+        async with httpx.AsyncClient(timeout=30.0) as client:
+            log.info(
+                "Sending webhook",
+                url=room.webhook_url,
+                payload_size=len(payload_bytes),
+            )
+
+            response = await client.post(
+                room.webhook_url,
+                content=payload_bytes,
+                headers=headers,
+            )
+
+            response.raise_for_status()
+
+            log.info(
+                "Webhook sent successfully",
+                status_code=response.status_code,
+                response_size=len(response.content),
+            )
+
+    except httpx.HTTPStatusError as e:
+        log.error(
+            "Webhook failed with HTTP error",
+            status_code=e.response.status_code,
+            response_text=e.response.text[:500],  # First 500 chars
+        )
+
+        # Don't retry on client errors (4xx)
+        if 400 <= e.response.status_code < 500:
+            log.error("Client error, not retrying")
+            return
+
+        # Retry on server errors (5xx)
+        raise self.retry(exc=e)
+
+    except (httpx.ConnectError, httpx.TimeoutException) as e:
+        # Retry on network errors
+        log.error("Webhook failed with connection error", error=str(e))
+        raise self.retry(exc=e)
+
+    except Exception as e:
+        # Retry on unexpected errors
+        log.exception("Unexpected error in webhook task", error=str(e))
+        raise self.retry(exc=e)
+
+
+async def test_webhook(room_id: str) -> dict:
+    """
+    Test webhook configuration by sending a sample payload.
+    Returns immediately with success/failure status.
+    This is the shared implementation used by both the API endpoint and Celery task.
+    """
+    try:
+        room = await rooms_controller.get_by_id(room_id)
+        if not room:
+            return {"success": False, "error": "Room not found"}
+
+        if not room.webhook_url:
+            return {"success": False, "error": "No webhook URL configured"}
+
+        now = (datetime.now(timezone.utc).isoformat(),)
+        payload_data = {
+            "event": "test",
+            "event_id": uuid.uuid4().hex,
+            "timestamp": now,
+            "message": "This is a test webhook from Reflector",
+            "room": {
+                "id": room.id,
+                "name": room.name,
+            },
+        }
+
+        payload_json = json.dumps(payload_data, separators=(",", ":"))
+        payload_bytes = payload_json.encode("utf-8")
+
+        # Generate headers with signature
+        headers = {
+            "Content-Type": "application/json",
+            "User-Agent": "Reflector-Webhook/1.0",
+            "X-Webhook-Event": "test",
+        }
+
+        if room.webhook_secret:
+            timestamp = str(int(datetime.now(timezone.utc).timestamp()))
+            signature = generate_webhook_signature(
+                payload_bytes, room.webhook_secret, timestamp
+            )
+            headers["X-Webhook-Signature"] = f"t={timestamp},v1={signature}"
+
+        # Send test webhook with short timeout
+        async with httpx.AsyncClient(timeout=10.0) as client:
+            response = await client.post(
+                room.webhook_url,
+                content=payload_bytes,
+                headers=headers,
+            )
+
+            return {
+                "success": response.is_success,
+                "status_code": response.status_code,
+                "message": f"Webhook test {'successful' if response.is_success else 'failed'}",
+                "response_preview": response.text if response.text else None,
+            }
+
+    except httpx.TimeoutException:
+        return {
+            "success": False,
+            "error": "Webhook request timed out (10 seconds)",
+        }
+    except httpx.ConnectError as e:
+        return {
+            "success": False,
+            "error": f"Could not connect to webhook URL: {str(e)}",
+        }
+    except Exception as e:
+        return {
+            "success": False,
+            "error": f"Unexpected error: {str(e)}",
+        }

server/run_jibri_tests.py

@@ -0,0 +1,49 @@
+#!/usr/bin/env python
+"""Simple test runner for Jibri tests that doesn't require Docker."""
+
+import os
+import sys
+
+sys.path.insert(0, os.path.dirname(os.path.abspath(__file__)))
+
+# Import test functions after path is set
+exec(open("tests/test_jibri_events.py").read(), globals())
+
+
+def run_tests():
+    tests = [
+        ("test_parse_room_created_event", test_parse_room_created_event),
+        ("test_parse_participant_joined_event", test_parse_participant_joined_event),
+        (
+            "test_parse_unknown_event_returns_none",
+            test_parse_unknown_event_returns_none,
+        ),
+        (
+            "test_parse_events_file_with_complete_session",
+            test_parse_events_file_with_complete_session,
+        ),
+        ("test_parse_events_file_missing_file", test_parse_events_file_missing_file),
+    ]
+
+    passed = 0
+    failed = 0
+
+    for name, test_func in tests:
+        try:
+            test_func()
+            print(f"✓ {name}")
+            passed += 1
+        except AssertionError as e:
+            print(f"✗ {name}: {e}")
+            failed += 1
+        except Exception as e:
+            print(f"✗ {name}: Unexpected error: {e}")
+            failed += 1
+
+    print(f"\nResults: {passed} passed, {failed} failed")
+    return failed == 0
+
+
+if __name__ == "__main__":
+    success = run_tests()
+    sys.exit(0 if success else 1)

server/runserver.sh

@@ -2,7 +2,7 @@
 
 if [ "${ENTRYPOINT}" = "server" ]; then
     uv run alembic upgrade head
-    uv run -m reflector.app
+    uv run uvicorn reflector.app:app --host 0.0.0.0 --port 1250
 elif [ "${ENTRYPOINT}" = "worker" ]; then
     uv run celery -A reflector.worker.app worker --loglevel=info
 elif [ "${ENTRYPOINT}" = "beat" ]; then

server/tests/conftest.py

@@ -178,6 +178,63 @@ async def dummy_diarization():
         yield
 
 
+@pytest.fixture
+async def dummy_file_transcript():
+    from reflector.processors.file_transcript import FileTranscriptProcessor
+    from reflector.processors.types import Transcript, Word
+
+    class TestFileTranscriptProcessor(FileTranscriptProcessor):
+        async def _transcript(self, data):
+            return Transcript(
+                text="Hello world. How are you today?",
+                words=[
+                    Word(start=0.0, end=0.5, text="Hello", speaker=0),
+                    Word(start=0.5, end=0.6, text=" ", speaker=0),
+                    Word(start=0.6, end=1.0, text="world", speaker=0),
+                    Word(start=1.0, end=1.1, text=".", speaker=0),
+                    Word(start=1.1, end=1.2, text=" ", speaker=0),
+                    Word(start=1.2, end=1.5, text="How", speaker=0),
+                    Word(start=1.5, end=1.6, text=" ", speaker=0),
+                    Word(start=1.6, end=1.8, text="are", speaker=0),
+                    Word(start=1.8, end=1.9, text=" ", speaker=0),
+                    Word(start=1.9, end=2.1, text="you", speaker=0),
+                    Word(start=2.1, end=2.2, text=" ", speaker=0),
+                    Word(start=2.2, end=2.5, text="today", speaker=0),
+                    Word(start=2.5, end=2.6, text="?", speaker=0),
+                ],
+            )
+
+    with patch(
+        "reflector.processors.file_transcript_auto.FileTranscriptAutoProcessor.__new__"
+    ) as mock_auto:
+        mock_auto.return_value = TestFileTranscriptProcessor()
+        yield
+
+
+@pytest.fixture
+async def dummy_file_diarization():
+    from reflector.processors.file_diarization import (
+        FileDiarizationOutput,
+        FileDiarizationProcessor,
+    )
+    from reflector.processors.types import DiarizationSegment
+
+    class TestFileDiarizationProcessor(FileDiarizationProcessor):
+        async def _diarize(self, data):
+            return FileDiarizationOutput(
+                diarization=[
+                    DiarizationSegment(start=0.0, end=1.1, speaker=0),
+                    DiarizationSegment(start=1.2, end=2.6, speaker=1),
+                ]
+            )
+
+    with patch(
+        "reflector.processors.file_diarization_auto.FileDiarizationAutoProcessor.__new__"
+    ) as mock_auto:
+        mock_auto.return_value = TestFileDiarizationProcessor()
+        yield
+
+
 @pytest.fixture
 async def dummy_transcript_translator():
     from reflector.processors.transcript_translator import TranscriptTranslatorProcessor
@@ -238,9 +295,13 @@ async def dummy_storage():
     with (
         patch("reflector.storage.base.Storage.get_instance") as mock_storage,
         patch("reflector.storage.get_transcripts_storage") as mock_get_transcripts,
+        patch(
+            "reflector.pipelines.main_file_pipeline.get_transcripts_storage"
+        ) as mock_get_transcripts2,
     ):
         mock_storage.return_value = dummy
         mock_get_transcripts.return_value = dummy
+        mock_get_transcripts2.return_value = dummy
         yield
 
 
@@ -260,7 +321,10 @@ def celery_config():
 
 @pytest.fixture(scope="session")
 def celery_includes():
-    return ["reflector.pipelines.main_live_pipeline"]
+    return [
+        "reflector.pipelines.main_live_pipeline",
+        "reflector.pipelines.main_file_pipeline",
+    ]
 
 
 @pytest.fixture
@@ -302,7 +366,7 @@ async def fake_transcript_with_topics(tmpdir, client):
     transcript = await transcripts_controller.get_by_id(tid)
     assert transcript is not None
 
-    await transcripts_controller.update(transcript, {"status": "finished"})
+    await transcripts_controller.update(transcript, {"status": "ended"})
 
     # manually copy a file at the expected location
     audio_filename = transcript.audio_mp3_filename

server/tests/test_cleanup.py

@@ -0,0 +1,285 @@
+from datetime import datetime, timedelta, timezone
+from unittest.mock import AsyncMock, patch
+
+import pytest
+
+from reflector.db.recordings import Recording, recordings_controller
+from reflector.db.transcripts import SourceKind, transcripts_controller
+from reflector.worker.cleanup import cleanup_old_public_data
+
+
+@pytest.mark.asyncio
+async def test_cleanup_old_public_data_skips_when_not_public():
+    """Test that cleanup is skipped when PUBLIC_MODE is False."""
+    with patch("reflector.worker.cleanup.settings") as mock_settings:
+        mock_settings.PUBLIC_MODE = False
+
+        result = await cleanup_old_public_data()
+
+        # Should return early without doing anything
+        assert result is None
+
+
+@pytest.mark.asyncio
+async def test_cleanup_old_public_data_deletes_old_anonymous_transcripts():
+    """Test that old anonymous transcripts are deleted."""
+    # Create old and new anonymous transcripts
+    old_date = datetime.now(timezone.utc) - timedelta(days=8)
+    new_date = datetime.now(timezone.utc) - timedelta(days=2)
+
+    # Create old anonymous transcript (should be deleted)
+    old_transcript = await transcripts_controller.add(
+        name="Old Anonymous Transcript",
+        source_kind=SourceKind.FILE,
+        user_id=None,  # Anonymous
+    )
+    # Manually update created_at to be old
+    from reflector.db import get_database
+    from reflector.db.transcripts import transcripts
+
+    await get_database().execute(
+        transcripts.update()
+        .where(transcripts.c.id == old_transcript.id)
+        .values(created_at=old_date)
+    )
+
+    # Create new anonymous transcript (should NOT be deleted)
+    new_transcript = await transcripts_controller.add(
+        name="New Anonymous Transcript",
+        source_kind=SourceKind.FILE,
+        user_id=None,  # Anonymous
+    )
+
+    # Create old transcript with user (should NOT be deleted)
+    old_user_transcript = await transcripts_controller.add(
+        name="Old User Transcript",
+        source_kind=SourceKind.FILE,
+        user_id="user123",
+    )
+    await get_database().execute(
+        transcripts.update()
+        .where(transcripts.c.id == old_user_transcript.id)
+        .values(created_at=old_date)
+    )
+
+    with patch("reflector.worker.cleanup.settings") as mock_settings:
+        mock_settings.PUBLIC_MODE = True
+        mock_settings.PUBLIC_DATA_RETENTION_DAYS = 7
+
+        # Mock the storage deletion
+        with patch("reflector.db.transcripts.get_transcripts_storage") as mock_storage:
+            mock_storage.return_value.delete_file = AsyncMock()
+
+            result = await cleanup_old_public_data()
+
+    # Check results
+    assert result["transcripts_deleted"] == 1
+    assert result["errors"] == []
+
+    # Verify old anonymous transcript was deleted
+    assert await transcripts_controller.get_by_id(old_transcript.id) is None
+
+    # Verify new anonymous transcript still exists
+    assert await transcripts_controller.get_by_id(new_transcript.id) is not None
+
+    # Verify user transcript still exists
+    assert await transcripts_controller.get_by_id(old_user_transcript.id) is not None
+
+
+@pytest.mark.asyncio
+async def test_cleanup_deletes_associated_meeting_and_recording():
+    """Test that meetings and recordings associated with old transcripts are deleted."""
+    from reflector.db import get_database
+    from reflector.db.meetings import meetings
+    from reflector.db.transcripts import transcripts
+
+    old_date = datetime.now(timezone.utc) - timedelta(days=8)
+
+    # Create a meeting
+    meeting_id = "test-meeting-for-transcript"
+    await get_database().execute(
+        meetings.insert().values(
+            id=meeting_id,
+            room_name="Meeting with Transcript",
+            room_url="https://example.com/meeting",
+            host_room_url="https://example.com/meeting-host",
+            start_date=old_date,
+            end_date=old_date + timedelta(hours=1),
+            room_id=None,
+        )
+    )
+
+    # Create a recording
+    recording = await recordings_controller.create(
+        Recording(
+            bucket_name="test-bucket",
+            object_key="test-recording.mp4",
+            recorded_at=old_date,
+        )
+    )
+
+    # Create an old transcript with both meeting and recording
+    old_transcript = await transcripts_controller.add(
+        name="Old Transcript with Meeting and Recording",
+        source_kind=SourceKind.ROOM,
+        user_id=None,
+        meeting_id=meeting_id,
+        recording_id=recording.id,
+    )
+
+    # Update created_at to be old
+    await get_database().execute(
+        transcripts.update()
+        .where(transcripts.c.id == old_transcript.id)
+        .values(created_at=old_date)
+    )
+
+    with patch("reflector.worker.cleanup.settings") as mock_settings:
+        mock_settings.PUBLIC_MODE = True
+        mock_settings.PUBLIC_DATA_RETENTION_DAYS = 7
+
+        # Mock storage deletion
+        with patch("reflector.db.transcripts.get_transcripts_storage") as mock_storage:
+            mock_storage.return_value.delete_file = AsyncMock()
+            with patch(
+                "reflector.worker.cleanup.get_recordings_storage"
+            ) as mock_rec_storage:
+                mock_rec_storage.return_value.delete_file = AsyncMock()
+
+                result = await cleanup_old_public_data()
+
+    # Check results
+    assert result["transcripts_deleted"] == 1
+    assert result["meetings_deleted"] == 1
+    assert result["recordings_deleted"] == 1
+    assert result["errors"] == []
+
+    # Verify transcript was deleted
+    assert await transcripts_controller.get_by_id(old_transcript.id) is None
+
+    # Verify meeting was deleted
+    query = meetings.select().where(meetings.c.id == meeting_id)
+    meeting_result = await get_database().fetch_one(query)
+    assert meeting_result is None
+
+    # Verify recording was deleted
+    assert await recordings_controller.get_by_id(recording.id) is None
+
+
+@pytest.mark.asyncio
+async def test_cleanup_handles_errors_gracefully():
+    """Test that cleanup continues even when individual deletions fail."""
+    old_date = datetime.now(timezone.utc) - timedelta(days=8)
+
+    # Create multiple old transcripts
+    transcript1 = await transcripts_controller.add(
+        name="Transcript 1",
+        source_kind=SourceKind.FILE,
+        user_id=None,
+    )
+    transcript2 = await transcripts_controller.add(
+        name="Transcript 2",
+        source_kind=SourceKind.FILE,
+        user_id=None,
+    )
+
+    # Update created_at to be old
+    from reflector.db import get_database
+    from reflector.db.transcripts import transcripts
+
+    for t_id in [transcript1.id, transcript2.id]:
+        await get_database().execute(
+            transcripts.update()
+            .where(transcripts.c.id == t_id)
+            .values(created_at=old_date)
+        )
+
+    with patch("reflector.worker.cleanup.settings") as mock_settings:
+        mock_settings.PUBLIC_MODE = True
+        mock_settings.PUBLIC_DATA_RETENTION_DAYS = 7
+
+        # Mock remove_by_id to fail for the first transcript
+        original_remove = transcripts_controller.remove_by_id
+        call_count = 0
+
+        async def mock_remove_by_id(transcript_id, user_id=None):
+            nonlocal call_count
+            call_count += 1
+            if call_count == 1:
+                raise Exception("Simulated deletion error")
+            return await original_remove(transcript_id, user_id)
+
+        with patch.object(
+            transcripts_controller, "remove_by_id", side_effect=mock_remove_by_id
+        ):
+            result = await cleanup_old_public_data()
+
+    # Should have one successful deletion and one error
+    assert result["transcripts_deleted"] == 1
+    assert len(result["errors"]) == 1
+    assert "Failed to delete transcript" in result["errors"][0]
+
+
+@pytest.mark.asyncio
+async def test_meeting_consent_cascade_delete():
+    """Test that meeting_consent records are automatically deleted when meeting is deleted."""
+    from reflector.db import get_database
+    from reflector.db.meetings import (
+        meeting_consent,
+        meeting_consent_controller,
+        meetings,
+    )
+
+    # Create a meeting
+    meeting_id = "test-cascade-meeting"
+    await get_database().execute(
+        meetings.insert().values(
+            id=meeting_id,
+            room_name="Test Meeting for CASCADE",
+            room_url="https://example.com/cascade-test",
+            host_room_url="https://example.com/cascade-test-host",
+            start_date=datetime.now(timezone.utc),
+            end_date=datetime.now(timezone.utc) + timedelta(hours=1),
+            room_id=None,
+        )
+    )
+
+    # Create consent records for this meeting
+    consent1_id = "consent-1"
+    consent2_id = "consent-2"
+
+    await get_database().execute(
+        meeting_consent.insert().values(
+            id=consent1_id,
+            meeting_id=meeting_id,
+            user_id="user1",
+            consent_given=True,
+            consent_timestamp=datetime.now(timezone.utc),
+        )
+    )
+
+    await get_database().execute(
+        meeting_consent.insert().values(
+            id=consent2_id,
+            meeting_id=meeting_id,
+            user_id="user2",
+            consent_given=False,
+            consent_timestamp=datetime.now(timezone.utc),
+        )
+    )
+
+    # Verify consent records exist
+    consents = await meeting_consent_controller.get_by_meeting_id(meeting_id)
+    assert len(consents) == 2
+
+    # Delete the meeting
+    await get_database().execute(meetings.delete().where(meetings.c.id == meeting_id))
+
+    # Verify meeting is deleted
+    query = meetings.select().where(meetings.c.id == meeting_id)
+    result = await get_database().fetch_one(query)
+    assert result is None
+
+    # Verify consent records are automatically deleted (CASCADE DELETE)
+    consents_after = await meeting_consent_controller.get_by_meeting_id(meeting_id)
+    assert len(consents_after) == 0

server/tests/test_gpu_modal_transcript.py

@@ -272,6 +272,9 @@ class TestGPUModalTranscript:
                 for f in temp_files:
                     Path(f).unlink(missing_ok=True)
 
+    @pytest.mark.skipif(
+        not "parakeet" in get_model_name(), reason="Parakeet only supports English"
+    )
     def test_transcriptions_error_handling(self):
         """Test error handling for invalid requests."""
         url = get_modal_transcript_url()

server/tests/test_jibri_events.py

@@ -0,0 +1,122 @@
+import json
+import tempfile
+from pathlib import Path
+
+from reflector.jibri_events import (
+    JitsiEventParser,
+    ParticipantJoinedEvent,
+    RoomCreatedEvent,
+)
+
+
+def test_parse_room_created_event():
+    parser = JitsiEventParser()
+    event_data = {
+        "type": "room_created",
+        "timestamp": 1234567890,
+        "room_name": "TestRoom",
+        "room_jid": "testroom@conference.meet.jitsi",
+        "meeting_url": "https://meet.jitsi/TestRoom",
+    }
+
+    event = parser.parse_event(event_data)
+
+    assert isinstance(event, RoomCreatedEvent)
+    assert event.room_name == "TestRoom"
+    assert event.meeting_url == "https://meet.jitsi/TestRoom"
+
+
+def test_parse_participant_joined_event():
+    parser = JitsiEventParser()
+    event_data = {
+        "type": "participant_joined",
+        "timestamp": 1234567891,
+        "room_name": "TestRoom",
+        "participant": {
+            "jid": "user1@meet.jitsi/resource",
+            "nick": "John Doe",
+            "id": "user1@meet.jitsi",
+            "is_moderator": False,
+        },
+    }
+
+    event = parser.parse_event(event_data)
+
+    assert isinstance(event, ParticipantJoinedEvent)
+    assert event.participant.nick == "John Doe"
+    assert event.participant.is_moderator is False
+
+
+def test_parse_unknown_event_returns_none():
+    parser = JitsiEventParser()
+    event_data = {"type": "unknown_event", "timestamp": 1234567890}
+
+    event = parser.parse_event(event_data)
+    assert event is None
+
+
+def test_parse_events_file_with_complete_session():
+    parser = JitsiEventParser()
+
+    with tempfile.TemporaryDirectory() as tmpdir:
+        events_file = Path(tmpdir) / "events.jsonl"
+
+        events = [
+            {
+                "type": "room_created",
+                "timestamp": 1234567890,
+                "room_name": "TestRoom",
+                "room_jid": "testroom@conference.meet.jitsi",
+                "meeting_url": "https://meet.jitsi/TestRoom",
+            },
+            {
+                "type": "participant_joined",
+                "timestamp": 1234567892,
+                "room_name": "TestRoom",
+                "participant": {
+                    "jid": "user1@meet.jitsi/resource",
+                    "nick": "John Doe",
+                    "id": "user1@meet.jitsi",
+                    "is_moderator": False,
+                },
+            },
+            {
+                "type": "speaker_active",
+                "timestamp": 1234567895,
+                "room_name": "TestRoom",
+                "speaker_jid": "user1@meet.jitsi",
+                "speaker_nick": "John Doe",
+                "duration": 10,
+            },
+            {
+                "type": "participant_left",
+                "timestamp": 1234567920,
+                "room_name": "TestRoom",
+                "participant": {
+                    "jid": "user1@meet.jitsi/resource",
+                    "duration_seconds": 28,
+                },
+            },
+        ]
+
+        with open(events_file, "w") as f:
+            for event in events:
+                f.write(json.dumps(event) + "\n")
+
+        metadata = parser.parse_events_file(tmpdir)
+
+        assert metadata["room"]["name"] == "TestRoom"
+        assert metadata["room"]["meeting_url"] == "https://meet.jitsi/TestRoom"
+        assert len(metadata["participants"]) == 1
+        assert metadata["event_count"] == 4
+
+
+def test_parse_events_file_missing_file():
+    parser = JitsiEventParser()
+
+    with tempfile.TemporaryDirectory() as tmpdir:
+        metadata = parser.parse_events_file(tmpdir)
+
+        assert metadata["room"]["name"] == ""
+        assert len(metadata["participants"]) == 0
+        assert metadata["event_count"] == 0

server/tests/test_jibri_webhook.py

@@ -0,0 +1,254 @@
+import json
+import tempfile
+from pathlib import Path
+from unittest.mock import Mock, patch
+
+import pytest
+from fastapi.testclient import TestClient
+from sqlalchemy.orm import Session
+
+from reflector.api.jibri_webhook import router
+from reflector.models import Transcript
+
+
+@pytest.fixture
+def client():
+    from fastapi import FastAPI
+
+    app = FastAPI()
+    app.include_router(router)
+    return TestClient(app)
+
+
+@pytest.fixture
+def mock_db():
+    db = Mock(spec=Session)
+    db.add = Mock()
+    db.commit = Mock()
+    db.refresh = Mock()
+    return db
+
+
+@pytest.fixture
+def mock_settings():
+    with patch("reflector.api.jibri_webhook.settings") as mock:
+        mock.JIBRI_RECORDINGS_PATH = "/recordings"
+        yield mock
+
+
+@pytest.fixture
+def mock_pipeline():
+    with patch("reflector.api.jibri_webhook.TranscriptMainPipeline") as mock:
+        pipeline_instance = Mock()
+        pipeline_instance.create = Mock()
+        mock.return_value = pipeline_instance
+        yield mock
+
+
+class TestJibriWebhook:
+    def test_recording_ready_success_with_events(
+        self, client, mock_db, mock_settings, mock_pipeline
+    ):
+        with tempfile.TemporaryDirectory() as tmpdir:
+            mock_settings.JIBRI_RECORDINGS_PATH = tmpdir
+
+            # Create recording directory and files
+            session_id = "test-session-123"
+            recording_dir = Path(tmpdir) / session_id
+            recording_dir.mkdir()
+
+            recording_file = recording_dir / "recording.mp4"
+            recording_file.write_text("fake video content")
+
+            events_file = recording_dir / "events.jsonl"
+            events = [
+                {
+                    "type": "room_created",
+                    "timestamp": 1234567890,
+                    "room_name": "TestRoom",
+                    "room_jid": "testroom@conference.meet.jitsi",
+                    "meeting_url": "https://meet.jitsi/TestRoom",
+                },
+                {
+                    "type": "participant_joined",
+                    "timestamp": 1234567892,
+                    "room_name": "TestRoom",
+                    "participant": {
+                        "jid": "user1@meet.jitsi/resource",
+                        "nick": "John Doe",
+                        "id": "user1@meet.jitsi",
+                        "is_moderator": False,
+                    },
+                },
+            ]
+
+            with open(events_file, "w") as f:
+                for event in events:
+                    f.write(json.dumps(event) + "\n")
+
+            # Mock database dependency
+            with patch("reflector.api.jibri_webhook.get_db") as mock_get_db:
+                mock_get_db.return_value = mock_db
+
+                response = client.post(
+                    "/api/v1/jibri/recording-ready",
+                    json={
+                        "session_id": session_id,
+                        "path": session_id,
+                        "meeting_url": "https://meet.jitsi/TestRoom",
+                    },
+                )
+
+            assert response.status_code == 200
+            data = response.json()
+
+            assert data["status"] == "accepted"
+            assert data["session_id"] == session_id
+            assert data["events_found"] is True
+            assert data["participant_count"] == 1
+
+            # Verify transcript was created
+            mock_db.add.assert_called_once()
+            transcript_arg = mock_db.add.call_args[0][0]
+            assert isinstance(transcript_arg, Transcript)
+            assert "TestRoom" in transcript_arg.title
+            assert transcript_arg.metadata["jitsi"]["room"]["name"] == "TestRoom"
+
+            # Verify pipeline was triggered
+            mock_pipeline.return_value.create.assert_called_once()
+
+    def test_recording_ready_success_without_events(
+        self, client, mock_db, mock_settings, mock_pipeline
+    ):
+        with tempfile.TemporaryDirectory() as tmpdir:
+            mock_settings.JIBRI_RECORDINGS_PATH = tmpdir
+
+            session_id = "test-session-456"
+            recording_dir = Path(tmpdir) / session_id
+            recording_dir.mkdir()
+
+            recording_file = recording_dir / "recording.mp4"
+            recording_file.write_text("fake video content")
+
+            with patch("reflector.api.jibri_webhook.get_db") as mock_get_db:
+                mock_get_db.return_value = mock_db
+
+                response = client.post(
+                    "/api/v1/jibri/recording-ready",
+                    json={
+                        "session_id": session_id,
+                        "path": session_id,
+                        "meeting_url": "https://meet.jitsi/NoEventsRoom",
+                    },
+                )
+
+            assert response.status_code == 200
+            data = response.json()
+
+            assert data["status"] == "accepted"
+            assert data["events_found"] is False
+            assert data["participant_count"] == 0
+
+            # Verify transcript was created with minimal metadata
+            mock_db.add.assert_called_once()
+            transcript_arg = mock_db.add.call_args[0][0]
+            assert transcript_arg.metadata["jitsi"]["participants"] == []
+
+    def test_recording_ready_path_not_found(self, client, mock_settings):
+        with tempfile.TemporaryDirectory() as tmpdir:
+            mock_settings.JIBRI_RECORDINGS_PATH = tmpdir
+
+            response = client.post(
+                "/api/v1/jibri/recording-ready",
+                json={
+                    "session_id": "nonexistent",
+                    "path": "nonexistent",
+                    "meeting_url": "https://meet.jitsi/Test",
+                },
+            )
+
+            assert response.status_code == 404
+            assert "Recording path not found" in response.json()["detail"]
+
+    def test_recording_ready_recording_file_not_found(self, client, mock_settings):
+        with tempfile.TemporaryDirectory() as tmpdir:
+            mock_settings.JIBRI_RECORDINGS_PATH = tmpdir
+
+            session_id = "test-no-recording"
+            recording_dir = Path(tmpdir) / session_id
+            recording_dir.mkdir()
+
+            # No recording.mp4 file created
+
+            response = client.post(
+                "/api/v1/jibri/recording-ready",
+                json={
+                    "session_id": session_id,
+                    "path": session_id,
+                    "meeting_url": "https://meet.jitsi/Test",
+                },
+            )
+
+            assert response.status_code == 404
+            assert "Recording file not found" in response.json()["detail"]
+
+    def test_recording_ready_with_relative_path(
+        self, client, mock_db, mock_settings, mock_pipeline
+    ):
+        with tempfile.TemporaryDirectory() as tmpdir:
+            mock_settings.JIBRI_RECORDINGS_PATH = tmpdir
+
+            # Create nested directory structure
+            session_id = "2024/01/15/test-session"
+            recording_dir = Path(tmpdir) / session_id
+            recording_dir.mkdir(parents=True)
+
+            recording_file = recording_dir / "recording.mp4"
+            recording_file.write_text("fake video content")
+
+            with patch("reflector.api.jibri_webhook.get_db") as mock_get_db:
+                mock_get_db.return_value = mock_db
+
+                response = client.post(
+                    "/api/v1/jibri/recording-ready",
+                    json={
+                        "session_id": "test-session",
+                        "path": session_id,  # Relative path with subdirectories
+                        "meeting_url": "https://meet.jitsi/Test",
+                    },
+                )
+
+            assert response.status_code == 200
+            data = response.json()
+            assert data["status"] == "accepted"
+
+    def test_recording_ready_empty_meeting_url(
+        self, client, mock_db, mock_settings, mock_pipeline
+    ):
+        with tempfile.TemporaryDirectory() as tmpdir:
+            mock_settings.JIBRI_RECORDINGS_PATH = tmpdir
+
+            session_id = "test-session"
+            recording_dir = Path(tmpdir) / session_id
+            recording_dir.mkdir()
+
+            recording_file = recording_dir / "recording.mp4"
+            recording_file.write_text("fake video content")
+
+            with patch("reflector.api.jibri_webhook.get_db") as mock_get_db:
+                mock_get_db.return_value = mock_db
+
+                response = client.post(
+                    "/api/v1/jibri/recording-ready",
+                    json={
+                        "session_id": session_id,
+                        "path": session_id,
+                        "meeting_url": "",
+                    },
+                )
+
+            assert response.status_code == 200
+
+            # Verify fallback URL was used
+            transcript_arg = mock_db.add.call_args[0][0]
+            assert transcript_arg.source_url == f"jitsi://{session_id}"

server/tests/test_processors_pipeline.py

@@ -1,61 +0,0 @@
-import pytest
-
-
-@pytest.mark.asyncio
-@pytest.mark.parametrize("enable_diarization", [False, True])
-async def test_basic_process(
-    dummy_transcript,
-    dummy_llm,
-    dummy_processors,
-    enable_diarization,
-    dummy_diarization,
-):
-    # goal is to start the server, and send rtc audio to it
-    # validate the events received
-    from pathlib import Path
-
-    from reflector.settings import settings
-    from reflector.tools.process import process_audio_file
-
-    # LLM_BACKEND no longer exists in settings
-    # settings.LLM_BACKEND = "test"
-    settings.TRANSCRIPT_BACKEND = "whisper"
-
-    # event callback
-    marks = {}
-
-    async def event_callback(event):
-        if event.processor not in marks:
-            marks[event.processor] = 0
-        marks[event.processor] += 1
-
-    # invoke the process and capture events
-    path = Path(__file__).parent / "records" / "test_mathieu_hello.wav"
-
-    if enable_diarization:
-        # Test with diarization - may fail if pyannote.audio is not installed
-        try:
-            await process_audio_file(
-                path.as_posix(), event_callback, enable_diarization=True
-            )
-        except SystemExit:
-            pytest.skip("pyannote.audio not installed - skipping diarization test")
-    else:
-        # Test without diarization - should always work
-        await process_audio_file(
-            path.as_posix(), event_callback, enable_diarization=False
-        )
-
-    print(f"Diarization: {enable_diarization}, Marks: {marks}")
-
-    # validate the events
-    # Each processor should be called for each audio segment processed
-    # The final processors (Topic, Title, Summary) should be called once at the end
-    assert marks["TranscriptLinerProcessor"] > 0
-    assert marks["TranscriptTranslatorPassthroughProcessor"] > 0
-    assert marks["TranscriptTopicDetectorProcessor"] == 1
-    assert marks["TranscriptFinalSummaryProcessor"] == 1
-    assert marks["TranscriptFinalTitleProcessor"] == 1
-
-    if enable_diarization:
-        assert marks["TestAudioDiarizationProcessor"] == 1

server/tests/test_search.py

@@ -23,7 +23,7 @@ async def test_search_postgresql_only():
     assert results == []
     assert total == 0
 
-    params_empty = SearchParameters(query_text="")
+    params_empty = SearchParameters(query_text=None)
     results_empty, total_empty = await search_controller.search_transcripts(
         params_empty
     )
@@ -34,7 +34,7 @@ async def test_search_postgresql_only():
 @pytest.mark.asyncio
 async def test_search_with_empty_query():
     """Test that empty query returns all transcripts."""
-    params = SearchParameters(query_text="")
+    params = SearchParameters(query_text=None)
     results, total = await search_controller.search_transcripts(params)
 
     assert isinstance(results, list)
@@ -58,7 +58,7 @@ async def test_empty_transcript_title_only_match():
             "id": test_id,
             "name": "Empty Transcript",
             "title": "Empty Meeting",
-            "status": "completed",
+            "status": "ended",
             "locked": False,
             "duration": 0.0,
             "created_at": datetime.now(timezone.utc),
@@ -109,7 +109,7 @@ async def test_search_with_long_summary():
             "id": test_id,
             "name": "Test Long Summary",
             "title": "Regular Meeting",
-            "status": "completed",
+            "status": "ended",
             "locked": False,
             "duration": 1800.0,
             "created_at": datetime.now(timezone.utc),
@@ -165,7 +165,7 @@ async def test_postgresql_search_with_data():
             "id": test_id,
             "name": "Test Search Transcript",
             "title": "Engineering Planning Meeting Q4 2024",
-            "status": "completed",
+            "status": "ended",
             "locked": False,
             "duration": 1800.0,
             "created_at": datetime.now(timezone.utc),
@@ -221,7 +221,7 @@ We need to implement PostgreSQL tsvector for better performance.""",
         test_result = next((r for r in results if r.id == test_id), None)
         if test_result:
             assert test_result.title == "Engineering Planning Meeting Q4 2024"
-            assert test_result.status == "completed"
+            assert test_result.status == "ended"
             assert test_result.duration == 1800.0
             assert 0 <= test_result.rank <= 1, "Rank should be normalized to 0-1"
 
@@ -268,7 +268,7 @@ def mock_db_result():
         "title": "Test Transcript",
         "created_at": datetime(2024, 6, 15, tzinfo=timezone.utc),
         "duration": 3600.0,
-        "status": "completed",
+        "status": "ended",
         "user_id": "test-user",
         "room_id": "room1",
         "source_kind": SourceKind.LIVE,
@@ -433,7 +433,7 @@ class TestSearchResultModel:
             room_id="room-456",
             source_kind=SourceKind.ROOM,
             created_at=datetime(2024, 6, 15, tzinfo=timezone.utc),
-            status="completed",
+            status="ended",
             rank=0.85,
             duration=1800.5,
             search_snippets=["snippet 1", "snippet 2"],
@@ -443,7 +443,7 @@ class TestSearchResultModel:
         assert result.title == "Test Title"
         assert result.user_id == "user-123"
         assert result.room_id == "room-456"
-        assert result.status == "completed"
+        assert result.status == "ended"
         assert result.rank == 0.85
         assert result.duration == 1800.5
         assert len(result.search_snippets) == 2
@@ -474,7 +474,7 @@ class TestSearchResultModel:
             id="test-id",
             source_kind=SourceKind.LIVE,
             created_at=datetime(2024, 6, 15, 12, 30, 45, tzinfo=timezone.utc),
-            status="completed",
+            status="ended",
             rank=0.9,
             duration=None,
             search_snippets=[],

server/tests/test_search_long_summary.py

@@ -25,7 +25,7 @@ async def test_long_summary_snippet_prioritization():
             "id": test_id,
             "name": "Test Snippet Priority",
             "title": "Meeting About Projects",
-            "status": "completed",
+            "status": "ended",
             "locked": False,
             "duration": 1800.0,
             "created_at": datetime.now(timezone.utc),
@@ -106,7 +106,7 @@ async def test_long_summary_only_search():
             "id": test_id,
             "name": "Test Long Only",
             "title": "Standard Meeting",
-            "status": "completed",
+            "status": "ended",
             "locked": False,
             "duration": 1800.0,
             "created_at": datetime.now(timezone.utc),

server/tests/test_search_snippets.py

@@ -1,5 +1,7 @@
 """Unit tests for search snippet generation."""
 
+import pytest
+
 from reflector.db.search import (
     SnippetCandidate,
     SnippetGenerator,
@@ -512,11 +514,9 @@ data visualization and data storage"""
         )
         assert webvtt_count == 3
 
-        snippets_empty, count_empty = SnippetGenerator.combine_sources(
-            None, None, "data", max_total=3
-        )
-        assert snippets_empty == []
-        assert count_empty == 0
+        # combine_sources requires at least one source to be present
+        with pytest.raises(AssertionError, match="At least one source must be present"):
+            SnippetGenerator.combine_sources(None, None, "data", max_total=3)
 
     def test_edge_cases(self):
         """Test edge cases for the pure functions."""

server/tests/test_transcripts_audio_download.py

@@ -19,7 +19,7 @@ async def fake_transcript(tmpdir, client):
     transcript = await transcripts_controller.get_by_id(tid)
     assert transcript is not None
 
-    await transcripts_controller.update(transcript, {"status": "finished"})
+    await transcripts_controller.update(transcript, {"status": "ended"})
 
     # manually copy a file at the expected location
     audio_filename = transcript.audio_mp3_filename

server/tests/test_transcripts_process.py

@@ -29,10 +29,10 @@ async def client(app_lifespan):
 @pytest.mark.asyncio
 async def test_transcript_process(
     tmpdir,
-    whisper_transcript,
     dummy_llm,
     dummy_processors,
-    dummy_diarization,
+    dummy_file_transcript,
+    dummy_file_diarization,
     dummy_storage,
     client,
 ):
@@ -56,8 +56,8 @@ async def test_transcript_process(
     assert response.status_code == 200
     assert response.json()["status"] == "ok"
 
-    # wait for processing to finish (max 10 minutes)
-    timeout_seconds = 600  # 10 minutes
+    # wait for processing to finish (max 1 minute)
+    timeout_seconds = 60
     start_time = time.monotonic()
     while (time.monotonic() - start_time) < timeout_seconds:
         # fetch the transcript and check if it is ended
@@ -75,9 +75,10 @@ async def test_transcript_process(
     )
     assert response.status_code == 200
     assert response.json()["status"] == "ok"
+    await asyncio.sleep(2)
 
-    # wait for processing to finish (max 10 minutes)
-    timeout_seconds = 600  # 10 minutes
+    # wait for processing to finish (max 1 minute)
+    timeout_seconds = 60
     start_time = time.monotonic()
     while (time.monotonic() - start_time) < timeout_seconds:
         # fetch the transcript and check if it is ended
@@ -99,4 +100,4 @@ async def test_transcript_process(
     response = await client.get(f"/transcripts/{tid}/topics")
     assert response.status_code == 200
     assert len(response.json()) == 1
-    assert "want to share" in response.json()[0]["transcript"]
+    assert "Hello world. How are you today?" in response.json()[0]["transcript"]

server/tests/test_transcripts_upload.py

@@ -12,7 +12,8 @@ async def test_transcript_upload_file(
     tmpdir,
     dummy_llm,
     dummy_processors,
-    dummy_diarization,
+    dummy_file_transcript,
+    dummy_file_diarization,
     dummy_storage,
     client,
 ):
@@ -36,8 +37,8 @@ async def test_transcript_upload_file(
     assert response.status_code == 200
     assert response.json()["status"] == "ok"
 
-    # wait the processing to finish (max 10 minutes)
-    timeout_seconds = 600  # 10 minutes
+    # wait the processing to finish (max 1 minute)
+    timeout_seconds = 60
     start_time = time.monotonic()
     while (time.monotonic() - start_time) < timeout_seconds:
         # fetch the transcript and check if it is ended
@@ -47,7 +48,7 @@ async def test_transcript_upload_file(
             break
         await asyncio.sleep(1)
     else:
-        pytest.fail(f"Processing timed out after {timeout_seconds} seconds")
+        return pytest.fail(f"Processing timed out after {timeout_seconds} seconds")
 
     # check the transcript is ended
     transcript = resp.json()
@@ -59,4 +60,4 @@ async def test_transcript_upload_file(
     response = await client.get(f"/transcripts/{tid}/topics")
     assert response.status_code == 200
     assert len(response.json()) == 1
-    assert "want to share" in response.json()[0]["transcript"]
+    assert "Hello world. How are you today?" in response.json()[0]["transcript"]

server/tests/test_video_platforms.py

@@ -0,0 +1,768 @@
+"""Tests for video platform abstraction and Jitsi integration."""
+
+from datetime import datetime, timedelta, timezone
+from unittest.mock import Mock, patch
+
+import pytest
+from fastapi.testclient import TestClient
+
+from reflector.db.rooms import Room, VideoPlatform
+from reflector.video_platforms.base import (
+    MeetingData,
+    VideoPlatformClient,
+    VideoPlatformConfig,
+)
+from reflector.video_platforms.factory import (
+    create_platform_client,
+    get_platform_config,
+)
+from reflector.video_platforms.jitsi import JitsiClient
+from reflector.video_platforms.registry import (
+    get_available_platforms,
+    get_platform_client,
+    register_platform,
+)
+from reflector.video_platforms.whereby import WherebyClient
+
+
+class TestVideoPlatformBase:
+    """Test the video platform base classes and interfaces."""
+
+    def test_video_platform_config_creation(self):
+        """Test VideoPlatformConfig can be created with required fields."""
+        config = VideoPlatformConfig(
+            api_key="test-key",
+            webhook_secret="test-secret",
+            api_url="https://test.example.com",
+        )
+        assert config.api_key == "test-key"
+        assert config.webhook_secret == "test-secret"
+        assert config.api_url == "https://test.example.com"
+
+    def test_meeting_data_creation(self):
+        """Test MeetingData can be created with all fields."""
+        meeting_data = MeetingData(
+            meeting_id="test-123",
+            room_name="test-room",
+            room_url="https://test.com/room",
+            host_room_url="https://test.com/host",
+            platform=VideoPlatform.JITSI,
+            extra_data={"jwt": "token123"},
+        )
+        assert meeting_data.meeting_id == "test-123"
+        assert meeting_data.room_name == "test-room"
+        assert meeting_data.platform == VideoPlatform.JITSI
+        assert meeting_data.extra_data["jwt"] == "token123"
+
+
+class TestJitsiClient:
+    """Test JitsiClient implementation."""
+
+    def setup_method(self):
+        """Set up test fixtures."""
+        self.config = VideoPlatformConfig(
+            api_key="",  # Jitsi doesn't use API key
+            webhook_secret="test-webhook-secret",
+            api_url="https://meet.example.com",
+        )
+        self.client = JitsiClient(self.config)
+        self.test_room = Room(
+            id="test-room-id", name="test-room", user_id="test-user", platform="jitsi"
+        )
+
+    @patch("reflector.settings.settings.JITSI_JWT_SECRET", "test-secret-123")
+    @patch("reflector.settings.settings.JITSI_DOMAIN", "meet.example.com")
+    @patch("reflector.settings.settings.JITSI_JWT_ISSUER", "reflector")
+    @patch("reflector.settings.settings.JITSI_JWT_AUDIENCE", "jitsi")
+    def test_jwt_generation(self):
+        """Test JWT token generation with proper payload."""
+        exp_time = datetime.now(timezone.utc) + timedelta(hours=1)
+        jwt_token = self.client._generate_jwt(
+            room="test-room", moderator=True, exp=exp_time
+        )
+
+        # Verify token is generated
+        assert jwt_token is not None
+        assert len(jwt_token) > 50  # JWT tokens are quite long
+        assert jwt_token.count(".") == 2  # JWT has 3 parts separated by dots
+
+    @patch("reflector.settings.settings.JITSI_JWT_SECRET", None)
+    def test_jwt_generation_without_secret_fails(self):
+        """Test JWT generation fails without secret."""
+        exp_time = datetime.now(timezone.utc) + timedelta(hours=1)
+
+        with pytest.raises(ValueError, match="JITSI_JWT_SECRET is required"):
+            self.client._generate_jwt(room="test-room", moderator=False, exp=exp_time)
+
+    @patch(
+        "reflector.video_platforms.jitsi.client.generate_uuid4",
+        return_value="test-uuid-123",
+    )
+    @patch("reflector.settings.settings.JITSI_JWT_SECRET", "test-secret-123")
+    @patch("reflector.settings.settings.JITSI_DOMAIN", "meet.example.com")
+    @patch("reflector.settings.settings.JITSI_JWT_ISSUER", "reflector")
+    @patch("reflector.settings.settings.JITSI_JWT_AUDIENCE", "jitsi")
+    async def test_create_meeting(self, mock_uuid):
+        """Test meeting creation with JWT tokens."""
+        end_date = datetime.now(timezone.utc) + timedelta(hours=2)
+
+        meeting_data = await self.client.create_meeting(
+            room_name_prefix="test", end_date=end_date, room=self.test_room
+        )
+
+        # Verify meeting data structure
+        assert meeting_data.meeting_id == "test-uuid-123"
+        assert meeting_data.platform == VideoPlatform.JITSI
+        assert "reflector-test-room" in meeting_data.room_name
+        assert "meet.example.com" in meeting_data.room_url
+        assert "jwt=" in meeting_data.room_url
+        assert "jwt=" in meeting_data.host_room_url
+
+        # Verify extra data contains JWT tokens
+        assert "user_jwt" in meeting_data.extra_data
+        assert "host_jwt" in meeting_data.extra_data
+        assert "domain" in meeting_data.extra_data
+
+    async def test_get_room_sessions(self):
+        """Test room sessions retrieval (mock implementation)."""
+        sessions = await self.client.get_room_sessions("test-room")
+
+        assert "roomName" in sessions
+        assert "sessions" in sessions
+        assert sessions["roomName"] == "test-room"
+        assert len(sessions["sessions"]) > 0
+        assert sessions["sessions"][0]["isActive"] is True
+
+    async def test_delete_room(self):
+        """Test room deletion (no-op for Jitsi)."""
+        result = await self.client.delete_room("test-room")
+        assert result is True
+
+    async def test_upload_logo(self):
+        """Test logo upload (no-op for Jitsi)."""
+        result = await self.client.upload_logo("test-room", "logo.png")
+        assert result is True
+
+    def test_verify_webhook_signature_valid(self):
+        """Test webhook signature verification with valid signature."""
+        body = b'{"event": "test"}'
+        # Generate expected signature
+        import hmac
+        from hashlib import sha256
+
+        expected_signature = hmac.new(
+            self.config.webhook_secret.encode(), body, sha256
+        ).hexdigest()
+
+        result = self.client.verify_webhook_signature(body, expected_signature)
+        assert result is True
+
+    def test_verify_webhook_signature_invalid(self):
+        """Test webhook signature verification with invalid signature."""
+        body = b'{"event": "test"}'
+        invalid_signature = "invalid-signature"
+
+        result = self.client.verify_webhook_signature(body, invalid_signature)
+        assert result is False
+
+    def test_verify_webhook_signature_no_secret(self):
+        """Test webhook signature verification without secret."""
+        config = VideoPlatformConfig(
+            api_key="", webhook_secret="", api_url="https://meet.example.com"
+        )
+        client = JitsiClient(config)
+
+        result = client.verify_webhook_signature(b'{"event": "test"}', "signature")
+        assert result is False
+
+
+class TestWherebyClient:
+    """Test WherebyClient implementation."""
+
+    def setup_method(self):
+        """Set up test fixtures."""
+        self.config = VideoPlatformConfig(
+            api_key="test-whereby-api-key",
+            webhook_secret="test-whereby-webhook-secret",
+            api_url="https://api.whereby.dev",
+            s3_bucket="test-recordings-bucket",
+            aws_access_key_id="test-access-key",
+            aws_access_key_secret="test-access-secret",
+        )
+        self.client = WherebyClient(self.config)
+        self.test_room = Room(
+            id="test-room-id",
+            name="test-room",
+            user_id="test-user",
+            platform=VideoPlatform.WHEREBY,
+        )
+
+    @patch("httpx.AsyncClient")
+    async def test_create_meeting(self, mock_client_class):
+        """Test Whereby meeting creation."""
+        # Mock the HTTP response
+        mock_client = mock_client_class.return_value.__aenter__.return_value
+        mock_response = Mock()
+        mock_response.json.return_value = {
+            "meetingId": "whereby-meeting-123",
+            "roomName": "whereby-room-456",
+            "roomUrl": "https://whereby.com/room",
+            "hostRoomUrl": "https://whereby.com/host-room",
+            "startDate": "2025-01-15T10:00:00.000Z",
+            "endDate": "2025-01-15T18:00:00.000Z",
+        }
+        mock_response.raise_for_status.return_value = None
+        mock_client.post.return_value = mock_response
+
+        end_date = datetime.now(timezone.utc) + timedelta(hours=2)
+
+        meeting_data = await self.client.create_meeting(
+            room_name_prefix="test", end_date=end_date, room=self.test_room
+        )
+
+        # Verify meeting data structure
+        assert meeting_data.meeting_id == "whereby-meeting-123"
+        assert meeting_data.room_name == "whereby-room-456"
+        assert meeting_data.platform == VideoPlatform.WHEREBY
+        assert "whereby.com" in meeting_data.room_url
+        assert "whereby.com" in meeting_data.host_room_url
+
+        # Verify HTTP call was made with correct parameters
+        mock_client.post.assert_called_once()
+        call_args = mock_client.post.call_args
+        assert "whereby.dev" in call_args[0][0]  # URL
+        assert "Bearer test-whereby-api-key" in call_args[1]["headers"]["Authorization"]
+
+    @patch("httpx.AsyncClient")
+    async def test_get_room_sessions(self, mock_client_class):
+        """Test Whereby room sessions retrieval."""
+        mock_client = mock_client_class.return_value.__aenter__.return_value
+        mock_response = Mock()
+        mock_response.json.return_value = {
+            "sessions": [
+                {
+                    "id": "session-123",
+                    "startTime": "2025-01-15T10:00:00Z",
+                    "participants": [],
+                }
+            ]
+        }
+        mock_response.raise_for_status.return_value = None
+        mock_client.get.return_value = mock_response
+
+        sessions = await self.client.get_room_sessions("test-room")
+
+        assert "sessions" in sessions
+        assert len(sessions["sessions"]) == 1
+        assert sessions["sessions"][0]["id"] == "session-123"
+
+        # Verify HTTP call
+        mock_client.get.assert_called_once()
+
+    async def test_delete_room(self):
+        """Test room deletion (no-op for Whereby)."""
+        result = await self.client.delete_room("test-room")
+        assert result is True
+
+    @patch("httpx.AsyncClient")
+    async def test_upload_logo_success(self, mock_client_class):
+        """Test logo upload success."""
+        mock_client = mock_client_class.return_value.__aenter__.return_value
+        mock_response = Mock()
+        mock_response.raise_for_status.return_value = None
+        mock_client.put.return_value = mock_response
+
+        # Create a temporary file for testing
+        import tempfile
+
+        with tempfile.NamedTemporaryFile(mode="w", suffix=".png", delete=False) as f:
+            f.write("fake logo content")
+            temp_file = f.name
+
+        result = await self.client.upload_logo("test-room", temp_file)
+        assert result is True
+
+        # Verify HTTP call
+        mock_client.put.assert_called_once()
+
+        # Cleanup
+        import os
+
+        os.unlink(temp_file)
+
+    @patch("httpx.AsyncClient")
+    async def test_upload_logo_failure(self, mock_client_class):
+        """Test logo upload handles HTTP errors gracefully."""
+        mock_client = mock_client_class.return_value.__aenter__.return_value
+        mock_client.put.side_effect = Exception("HTTP error")
+
+        result = await self.client.upload_logo("test-room", "logo.png")
+        assert result is False
+
+    def test_verify_webhook_signature_valid(self):
+        """Test Whereby webhook signature verification with valid signature."""
+        body = b'{"event": "test"}'
+        import hmac
+        from hashlib import sha256
+
+        expected_signature = hmac.new(
+            self.config.webhook_secret.encode(), body, sha256
+        ).hexdigest()
+
+        result = self.client.verify_webhook_signature(body, expected_signature)
+        assert result is True
+
+    def test_verify_webhook_signature_invalid(self):
+        """Test Whereby webhook signature verification with invalid signature."""
+        body = b'{"event": "test"}'
+        invalid_signature = "invalid-signature"
+
+        result = self.client.verify_webhook_signature(body, invalid_signature)
+        assert result is False
+
+
+class TestPlatformRegistry:
+    """Test platform registry functionality."""
+
+    def test_platform_registration(self):
+        """Test platform registration and retrieval."""
+
+        # Create mock client class
+        class MockClient(VideoPlatformClient):
+            async def create_meeting(self, room_name_prefix, end_date, room):
+                pass
+
+            async def get_room_sessions(self, room_name):
+                pass
+
+            async def delete_room(self, room_name):
+                pass
+
+            async def upload_logo(self, room_name, logo_path):
+                pass
+
+            def verify_webhook_signature(self, body, signature, timestamp=None):
+                pass
+
+        # Register mock platform
+        register_platform("test-platform", MockClient)
+
+        # Verify it's available
+        available = get_available_platforms()
+        assert "test-platform" in available
+
+        # Test client creation
+        config = VideoPlatformConfig(
+            api_key="test", webhook_secret="test", api_url="test"
+        )
+        client = get_platform_client("test-platform", config)
+        assert isinstance(client, MockClient)
+
+    def test_get_unknown_platform_raises_error(self):
+        """Test that requesting unknown platform raises error."""
+        config = VideoPlatformConfig(
+            api_key="test", webhook_secret="test", api_url="test"
+        )
+
+        with pytest.raises(ValueError, match="Unknown video platform: nonexistent"):
+            get_platform_client("nonexistent", config)
+
+    def test_builtin_platforms_registered(self):
+        """Test that built-in platforms are registered."""
+        available = get_available_platforms()
+        assert "jitsi" in available
+        assert "whereby" in available
+
+
+class TestPlatformFactory:
+    """Test platform factory functionality."""
+
+    @patch("reflector.settings.settings.JITSI_JWT_SECRET", "test-secret")
+    @patch("reflector.settings.settings.JITSI_WEBHOOK_SECRET", "webhook-secret")
+    @patch("reflector.settings.settings.JITSI_DOMAIN", "meet.example.com")
+    def test_get_jitsi_platform_config(self):
+        """Test Jitsi platform configuration."""
+        config = get_platform_config("jitsi")
+
+        assert config.api_key == ""  # Jitsi uses JWT, no API key
+        assert config.webhook_secret == "webhook-secret"
+        assert config.api_url == "https://meet.example.com"
+
+    @patch("reflector.settings.settings.WHEREBY_API_KEY", "whereby-key")
+    @patch("reflector.settings.settings.WHEREBY_WEBHOOK_SECRET", "whereby-secret")
+    @patch("reflector.settings.settings.WHEREBY_API_URL", "https://api.whereby.dev")
+    def test_get_whereby_platform_config(self):
+        """Test Whereby platform configuration."""
+        config = get_platform_config("whereby")
+
+        assert config.api_key == "whereby-key"
+        assert config.webhook_secret == "whereby-secret"
+        assert config.api_url == "https://api.whereby.dev"
+
+    def test_get_unknown_platform_config_raises_error(self):
+        """Test that unknown platform config raises error."""
+        with pytest.raises(ValueError, match="Unknown platform: nonexistent"):
+            get_platform_config("nonexistent")
+
+    def test_create_platform_client(self):
+        """Test platform client creation via factory."""
+        with patch(
+            "reflector.video_platforms.factory.get_platform_config"
+        ) as mock_config:
+            mock_config.return_value = VideoPlatformConfig(
+                api_key="",
+                webhook_secret="test-secret",
+                api_url="https://meet.example.com",
+            )
+
+            client = create_platform_client("jitsi")
+            assert isinstance(client, JitsiClient)
+
+    def test_create_jitsi_client_typing(self):
+        """Test that create_platform_client returns correctly typed JitsiClient."""
+        with patch(
+            "reflector.video_platforms.factory.get_platform_config"
+        ) as mock_config:
+            mock_config.return_value = VideoPlatformConfig(
+                api_key="",
+                webhook_secret="test-secret",
+                api_url="https://meet.example.com",
+            )
+
+            # The typing overload should ensure this returns JitsiClient
+            client = create_platform_client("jitsi")
+            assert isinstance(client, JitsiClient)
+            # Verify it has Jitsi-specific methods
+            assert hasattr(client, "_generate_jwt")
+
+    def test_create_whereby_client_typing(self):
+        """Test that create_platform_client returns correctly typed WherebyClient."""
+        with patch(
+            "reflector.video_platforms.factory.get_platform_config"
+        ) as mock_config:
+            mock_config.return_value = VideoPlatformConfig(
+                api_key="whereby-key",
+                webhook_secret="whereby-secret",
+                api_url="https://api.whereby.dev",
+            )
+
+            # The typing overload should ensure this returns WherebyClient
+            client = create_platform_client("whereby")
+            assert isinstance(client, WherebyClient)
+            # Verify it has Whereby-specific attributes
+            assert hasattr(client, "headers")
+            assert hasattr(client, "timeout")
+
+
+class TestWebhookEventStorage:
+    """Test webhook event storage functionality."""
+
+    def setup_method(self):
+        """Set up test fixtures."""
+        from reflector.app import app
+
+        self.client = TestClient(app)
+
+    @patch("reflector.db.meetings.meetings_controller.participant_joined")
+    @patch("reflector.db.meetings.meetings_controller.get_by_room_name")
+    @patch(
+        "reflector.video_platforms.jitsi.router.verify_jitsi_webhook_signature",
+        return_value=True,
+    )
+    def test_participant_joined_event_storage(
+        self, mock_verify, mock_get, mock_participant_joined
+    ):
+        """Test that participant joined events are stored correctly."""
+        # Mock meeting
+        mock_meeting = Mock()
+        mock_meeting.id = "test-meeting-id"
+        mock_meeting.num_clients = 1
+        mock_get.return_value = mock_meeting
+
+        payload = {
+            "event": "muc-occupant-joined",
+            "room": "test-room",
+            "timestamp": "2025-01-15T10:30:00.000Z",
+            "data": {"user_id": "test-user", "display_name": "John Doe"},
+        }
+
+        response = self.client.post(
+            "/v1/jitsi/events",
+            json=payload,
+            headers={"x-jitsi-signature": "valid-signature"},
+        )
+
+        assert response.status_code == 200
+        # Verify event was stored with correct data
+        mock_participant_joined.assert_called_once_with(
+            "test-meeting-id",
+            {
+                "timestamp": datetime.fromisoformat(
+                    "2025-01-15T10:30:00.000Z".replace("Z", "+00:00")
+                ),
+                "data": {"user_id": "test-user", "display_name": "John Doe"},
+            },
+        )
+
+    @patch("reflector.db.meetings.meetings_controller.recording_started")
+    @patch("reflector.db.meetings.meetings_controller.get_by_room_name")
+    @patch(
+        "reflector.video_platforms.jitsi.router.verify_jitsi_webhook_signature",
+        return_value=True,
+    )
+    def test_recording_started_event_storage(
+        self, mock_verify, mock_get, mock_recording_started
+    ):
+        """Test that recording started events are stored correctly."""
+        mock_meeting = Mock()
+        mock_meeting.id = "test-meeting-id"
+        mock_meeting.num_clients = 1
+        mock_get.return_value = mock_meeting
+
+        payload = {
+            "event": "jibri-recording-on",
+            "room": "test-room",
+            "timestamp": "2025-01-15T10:32:00.000Z",
+            "data": {"recording_id": "rec-123"},
+        }
+
+        response = self.client.post(
+            "/v1/jitsi/events",
+            json=payload,
+            headers={"x-jitsi-signature": "valid-signature"},
+        )
+
+        assert response.status_code == 200
+        mock_recording_started.assert_called_once_with(
+            "test-meeting-id",
+            {
+                "timestamp": datetime.fromisoformat(
+                    "2025-01-15T10:32:00.000Z".replace("Z", "+00:00")
+                ),
+                "data": {"recording_id": "rec-123"},
+            },
+        )
+
+    @patch("reflector.db.meetings.meetings_controller.add_event")
+    @patch("reflector.db.meetings.meetings_controller.get_by_room_name")
+    @patch(
+        "reflector.video_platforms.jitsi.router.verify_jitsi_webhook_signature",
+        return_value=True,
+    )
+    def test_recording_complete_event_storage(
+        self, mock_verify, mock_get, mock_add_event
+    ):
+        """Test that recording completion events are stored correctly."""
+        mock_meeting = Mock()
+        mock_meeting.id = "test-meeting-id"
+        mock_meeting.num_clients = 1
+        mock_get.return_value = mock_meeting
+
+        payload = {
+            "room_name": "test-room",
+            "recording_file": "/recordings/test.mp4",
+            "recording_status": "completed",
+            "timestamp": "2025-01-15T11:15:00.000Z",
+        }
+
+        response = self.client.post(
+            "/v1/jibri/recording-complete",
+            json=payload,
+            headers={"x-jitsi-signature": "valid-signature"},
+        )
+
+        assert response.status_code == 200
+        mock_add_event.assert_called_once_with(
+            "test-meeting-id",
+            "recording_completed",
+            {
+                "recording_file": "/recordings/test.mp4",
+                "recording_status": "completed",
+                "timestamp": datetime.fromisoformat(
+                    "2025-01-15T11:15:00.000Z".replace("Z", "+00:00")
+                ),
+            },
+        )
+
+
+class TestWebhookEndpoints:
+    """Test Jitsi webhook endpoints."""
+
+    def setup_method(self):
+        """Set up test client."""
+        from reflector.app import app
+
+        self.client = TestClient(app)
+
+    def test_health_endpoint(self):
+        """Test Jitsi health check endpoint."""
+        response = self.client.get("/v1/jitsi/health")
+
+        assert response.status_code == 200
+        data = response.json()
+        assert data["status"] == "ok"
+        assert data["service"] == "jitsi-webhooks"
+        assert "timestamp" in data
+        assert "webhook_secret_configured" in data
+
+    @patch(
+        "reflector.video_platforms.jitsi.router.verify_jitsi_webhook_signature",
+        return_value=True,
+    )
+    @patch("reflector.db.meetings.meetings_controller.get_by_room_name")
+    @patch("reflector.db.meetings.meetings_controller.participant_joined")
+    @patch("reflector.db.meetings.meetings_controller.update_meeting")
+    async def test_jitsi_events_webhook_join(
+        self, mock_update, mock_participant_joined, mock_get, mock_verify
+    ):
+        """Test participant join event webhook."""
+        # Mock meeting
+        mock_meeting = Mock()
+        mock_meeting.id = "test-meeting-id"
+        mock_meeting.num_clients = 1
+        mock_get.return_value = mock_meeting
+
+        payload = {
+            "event": "muc-occupant-joined",
+            "room": "test-room",
+            "timestamp": "2025-01-15T10:30:00.000Z",
+            "data": {},
+        }
+
+        response = self.client.post(
+            "/v1/jitsi/events",
+            json=payload,
+            headers={"x-jitsi-signature": "valid-signature"},
+        )
+
+        assert response.status_code == 200
+        data = response.json()
+        assert data["status"] == "ok"
+        assert data["event"] == "muc-occupant-joined"
+        assert data["room"] == "test-room"
+
+    @patch(
+        "reflector.video_platforms.jitsi.router.verify_jitsi_webhook_signature",
+        return_value=False,
+    )
+    async def test_jitsi_events_webhook_invalid_signature(self, mock_verify):
+        """Test webhook with invalid signature returns 401."""
+        payload = {
+            "event": "muc-occupant-joined",
+            "room": "test-room",
+            "timestamp": "2025-01-15T10:30:00.000Z",
+            "data": {},
+        }
+
+        response = self.client.post(
+            "/v1/jitsi/events",
+            json=payload,
+            headers={"x-jitsi-signature": "invalid-signature"},
+        )
+
+        assert response.status_code == 401
+        assert "Invalid webhook signature" in response.text
+
+    @patch(
+        "reflector.video_platforms.jitsi.router.verify_jitsi_webhook_signature",
+        return_value=True,
+    )
+    @patch(
+        "reflector.db.meetings.meetings_controller.get_by_room_name", return_value=None
+    )
+    async def test_jitsi_events_webhook_meeting_not_found(self, mock_get, mock_verify):
+        """Test webhook with nonexistent meeting returns 404."""
+        payload = {
+            "event": "muc-occupant-joined",
+            "room": "nonexistent-room",
+            "timestamp": "2025-01-15T10:30:00.000Z",
+            "data": {},
+        }
+
+        response = self.client.post(
+            "/v1/jitsi/events",
+            json=payload,
+            headers={"x-jitsi-signature": "valid-signature"},
+        )
+
+        assert response.status_code == 404
+        assert "Meeting not found" in response.text
+
+
+class TestRoomsPlatformIntegration:
+    """Test rooms endpoint integration with platform abstraction."""
+
+    def setup_method(self):
+        """Set up test client."""
+        from reflector.app import app
+
+        self.client = TestClient(app)
+
+    @patch("reflector.auth.current_user_optional")
+    @patch("reflector.db.rooms.rooms_controller.add")
+    def test_create_room_with_jitsi_platform(self, mock_add, mock_auth):
+        """Test room creation with Jitsi platform."""
+        from datetime import datetime, timezone
+
+        mock_auth.return_value = {"sub": "test-user"}
+
+        # Create a proper Room object for the mock return
+        from reflector.db.rooms import Room
+
+        mock_room = Room(
+            id="test-room-id",
+            name="test-jitsi-room",
+            user_id="test-user",
+            created_at=datetime.now(timezone.utc),
+            zulip_auto_post=False,
+            zulip_stream="",
+            zulip_topic="",
+            is_locked=False,
+            room_mode="normal",
+            recording_type="cloud",
+            recording_trigger="automatic-2nd-participant",
+            is_shared=False,
+            platform=VideoPlatform.JITSI,
+        )
+        mock_add.return_value = mock_room
+
+        payload = {
+            "name": "test-jitsi-room",
+            "platform": "jitsi",
+            "zulip_auto_post": False,
+            "zulip_stream": "",
+            "zulip_topic": "",
+            "is_locked": False,
+            "room_mode": "normal",
+            "recording_type": "cloud",
+            "recording_trigger": "automatic-2nd-participant",
+            "is_shared": False,
+            "webhook_url": "",
+            "webhook_secret": "",
+        }
+
+        response = self.client.post("/v1/rooms", json=payload)
+
+        # Verify the add method was called with platform parameter
+        mock_add.assert_called_once()
+        call_args = mock_add.call_args
+        assert call_args.kwargs["platform"] == "jitsi"
+        assert call_args.kwargs["name"] == "test-jitsi-room"
+        assert response.status_code == 200
+
+    def test_create_meeting_with_jitsi_platform_fallback(self):
+        """Test that meeting creation falls back to whereby when platform client unavailable."""
+        # This tests the fallback behavior in rooms.py when platform client returns None
+        # The actual platform integration test is covered in the unit tests above
+
+        # Just verify the endpoint exists and has the right structure
+        # More detailed integration testing would require a full test database setup
+        assert hasattr(self.client.app, "routes")
+
+        # Find the meeting creation route
+        meeting_routes = [
+            r
+            for r in self.client.app.routes
+            if hasattr(r, "path") and "meeting" in r.path
+        ]
+        assert len(meeting_routes) > 0

server/uv.lock

@@ -2706,6 +2706,15 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/0c/7f/113b16d55e8d2dd9143628eec39b138fd6c52f72dcd11b4dae4a3845da4d/pyinstrument-5.0.3-cp312-cp312-win_amd64.whl", hash = "sha256:88df7e3ab11604ae7cef1f576c097a08752bf8fc13c5755803bd3cd92f15aba3", size = 124314, upload-time = "2025-07-02T14:13:26.708Z" },
 ]
 
+[[package]]
+name = "pyjwt"
+version = "2.10.1"
+source = { registry = "https://pypi.org/simple" }
+sdist = { url = "https://files.pythonhosted.org/packages/e7/46/bd74733ff231675599650d3e47f361794b22ef3e3770998dda30d3b63726/pyjwt-2.10.1.tar.gz", hash = "sha256:3cc5772eb20009233caf06e9d8a0577824723b44e6648ee0a2aedb6cf9381953", size = 87785, upload-time = "2024-11-28T03:43:29.933Z" }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/61/ad/689f02752eeec26aed679477e80e632ef1b682313be70793d798c1d5fc8f/PyJWT-2.10.1-py3-none-any.whl", hash = "sha256:dcdd193e30abefd5debf142f9adfcdd2b58004e644f25406ffaebd50bd98dacb", size = 22997, upload-time = "2024-11-28T03:43:27.893Z" },
+]
+
 [[package]]
 name = "pylibsrtp"
 version = "0.12.0"
@@ -3136,6 +3145,7 @@ dependencies = [
     { name = "protobuf" },
     { name = "psycopg2-binary" },
     { name = "pydantic-settings" },
+    { name = "pyjwt" },
     { name = "pytest-env" },
     { name = "python-jose", extra = ["cryptography"] },
     { name = "python-multipart" },
@@ -3213,6 +3223,7 @@ requires-dist = [
     { name = "protobuf", specifier = ">=4.24.3" },
     { name = "psycopg2-binary", specifier = ">=2.9.10" },
     { name = "pydantic-settings", specifier = ">=2.0.2" },
+    { name = "pyjwt", specifier = ">=2.8.0" },
     { name = "pytest-env", specifier = ">=1.1.5" },
     { name = "python-jose", extras = ["cryptography"], specifier = ">=3.3.0" },
     { name = "python-multipart", specifier = ">=0.0.6" },
@@ -3954,8 +3965,8 @@ dependencies = [
     { name = "typing-extensions", marker = "platform_python_implementation != 'PyPy' and sys_platform == 'darwin'" },
 ]
 wheels = [
-    { url = "https://download.pytorch.org/whl/cpu/torch-2.8.0-cp311-none-macosx_11_0_arm64.whl", hash = "sha256:3d05017d19bc99741288e458888283a44b0ee881d53f05f72f8b1cfea8998122" },
-    { url = "https://download.pytorch.org/whl/cpu/torch-2.8.0-cp312-none-macosx_11_0_arm64.whl", hash = "sha256:a47b7986bee3f61ad217d8a8ce24605809ab425baf349f97de758815edd2ef54" },
+    { url = "https://download.pytorch.org/whl/cpu/torch-2.8.0-cp311-none-macosx_11_0_arm64.whl" },
+    { url = "https://download.pytorch.org/whl/cpu/torch-2.8.0-cp312-none-macosx_11_0_arm64.whl" },
 ]
 
 [[package]]
@@ -3980,16 +3991,16 @@ dependencies = [
     { name = "typing-extensions", marker = "platform_python_implementation == 'PyPy' or sys_platform != 'darwin'" },
 ]
 wheels = [
-    { url = "https://download.pytorch.org/whl/cpu/torch-2.8.0%2Bcpu-cp311-cp311-linux_s390x.whl", hash = "sha256:2bfc013dd6efdc8f8223a0241d3529af9f315dffefb53ffa3bf14d3f10127da6" },
-    { url = "https://download.pytorch.org/whl/cpu/torch-2.8.0%2Bcpu-cp311-cp311-manylinux_2_28_aarch64.whl", hash = "sha256:680129efdeeec3db5da3f88ee5d28c1b1e103b774aef40f9d638e2cce8f8d8d8" },
-    { url = "https://download.pytorch.org/whl/cpu/torch-2.8.0%2Bcpu-cp311-cp311-manylinux_2_28_x86_64.whl", hash = "sha256:cb06175284673a581dd91fb1965662ae4ecaba6e5c357aa0ea7bb8b84b6b7eeb" },
-    { url = "https://download.pytorch.org/whl/cpu/torch-2.8.0%2Bcpu-cp311-cp311-win_amd64.whl", hash = "sha256:7631ef49fbd38d382909525b83696dc12a55d68492ade4ace3883c62b9fc140f" },
-    { url = "https://download.pytorch.org/whl/cpu/torch-2.8.0%2Bcpu-cp311-cp311-win_arm64.whl", hash = "sha256:41e6fc5ec0914fcdce44ccf338b1d19a441b55cafdd741fd0bf1af3f9e4cfd14" },
-    { url = "https://download.pytorch.org/whl/cpu/torch-2.8.0%2Bcpu-cp312-cp312-linux_s390x.whl", hash = "sha256:0e34e276722ab7dd0dffa9e12fe2135a9b34a0e300c456ed7ad6430229404eb5" },
-    { url = "https://download.pytorch.org/whl/cpu/torch-2.8.0%2Bcpu-cp312-cp312-manylinux_2_28_aarch64.whl", hash = "sha256:610f600c102386e581327d5efc18c0d6edecb9820b4140d26163354a99cd800d" },
-    { url = "https://download.pytorch.org/whl/cpu/torch-2.8.0%2Bcpu-cp312-cp312-manylinux_2_28_x86_64.whl", hash = "sha256:cb9a8ba8137ab24e36bf1742cb79a1294bd374db570f09fc15a5e1318160db4e" },
-    { url = "https://download.pytorch.org/whl/cpu/torch-2.8.0%2Bcpu-cp312-cp312-win_amd64.whl", hash = "sha256:2be20b2c05a0cce10430cc25f32b689259640d273232b2de357c35729132256d" },
-    { url = "https://download.pytorch.org/whl/cpu/torch-2.8.0%2Bcpu-cp312-cp312-win_arm64.whl", hash = "sha256:99fc421a5d234580e45957a7b02effbf3e1c884a5dd077afc85352c77bf41434" },
+    { url = "https://download.pytorch.org/whl/cpu/torch-2.8.0%2Bcpu-cp311-cp311-linux_s390x.whl" },
+    { url = "https://download.pytorch.org/whl/cpu/torch-2.8.0%2Bcpu-cp311-cp311-manylinux_2_28_aarch64.whl" },
+    { url = "https://download.pytorch.org/whl/cpu/torch-2.8.0%2Bcpu-cp311-cp311-manylinux_2_28_x86_64.whl" },
+    { url = "https://download.pytorch.org/whl/cpu/torch-2.8.0%2Bcpu-cp311-cp311-win_amd64.whl" },
+    { url = "https://download.pytorch.org/whl/cpu/torch-2.8.0%2Bcpu-cp311-cp311-win_arm64.whl" },
+    { url = "https://download.pytorch.org/whl/cpu/torch-2.8.0%2Bcpu-cp312-cp312-linux_s390x.whl" },
+    { url = "https://download.pytorch.org/whl/cpu/torch-2.8.0%2Bcpu-cp312-cp312-manylinux_2_28_aarch64.whl" },
+    { url = "https://download.pytorch.org/whl/cpu/torch-2.8.0%2Bcpu-cp312-cp312-manylinux_2_28_x86_64.whl" },
+    { url = "https://download.pytorch.org/whl/cpu/torch-2.8.0%2Bcpu-cp312-cp312-win_amd64.whl" },
+    { url = "https://download.pytorch.org/whl/cpu/torch-2.8.0%2Bcpu-cp312-cp312-win_arm64.whl" },
 ]
 
 [[package]]

www/.env.example

@@ -0,0 +1,34 @@
+# Environment
+ENVIRONMENT=development
+NEXT_PUBLIC_ENV=development
+
+# Site Configuration
+NEXT_PUBLIC_SITE_URL=http://localhost:3000
+
+# Nextauth envs
+# not used in app code but in lib code
+NEXTAUTH_URL=http://localhost:3000
+NEXTAUTH_SECRET=your-nextauth-secret-here
+# / Nextauth envs
+
+# Authentication (Authentik OAuth/OIDC)
+AUTHENTIK_ISSUER=https://authentik.example.com/application/o/reflector
+AUTHENTIK_REFRESH_TOKEN_URL=https://authentik.example.com/application/o/token/
+AUTHENTIK_CLIENT_ID=your-client-id-here
+AUTHENTIK_CLIENT_SECRET=your-client-secret-here
+
+# Feature Flags
+# NEXT_PUBLIC_FEATURE_REQUIRE_LOGIN=true
+# NEXT_PUBLIC_FEATURE_PRIVACY=false
+# NEXT_PUBLIC_FEATURE_BROWSE=true
+# NEXT_PUBLIC_FEATURE_SEND_TO_ZULIP=true
+# NEXT_PUBLIC_FEATURE_ROOMS=true
+
+# API URLs
+NEXT_PUBLIC_API_URL=http://127.0.0.1:1250
+NEXT_PUBLIC_WEBSOCKET_URL=ws://127.0.0.1:1250
+NEXT_PUBLIC_AUTH_CALLBACK_URL=http://localhost:3000/auth-callback
+
+# Sentry
+# SENTRY_DSN=https://your-dsn@sentry.io/project-id
+# SENTRY_IGNORE_API_RESOLUTION_ERROR=1

www/.env.template

@@ -0,0 +1,46 @@
+# NextAuth configuration
+NEXTAUTH_SECRET="your-secret-key"
+NEXTAUTH_URL="http://localhost:3000/"
+
+# API configuration
+NEXT_PUBLIC_API_URL="http://127.0.0.1:1250"
+NEXT_PUBLIC_WEBSOCKET_URL="ws://127.0.0.1:1250"
+NEXT_PUBLIC_AUTH_CALLBACK_URL="http://localhost:3000/auth-callback"
+NEXT_PUBLIC_SITE_URL="http://localhost:3000/"
+
+# Environment
+NEXT_PUBLIC_ENV="development"
+ENVIRONMENT="development"
+
+# Video Platform Configuration
+# Options: "whereby" | "jitsi" (default: whereby)
+NEXT_PUBLIC_VIDEO_PLATFORM="whereby"
+
+# Features
+NEXT_PUBLIC_PROJECTOR_MODE="false"
+
+# Authentication providers (optional)
+# Authentik
+AUTHENTIK_CLIENT_ID=""
+AUTHENTIK_CLIENT_SECRET=""
+AUTHENTIK_ISSUER=""
+AUTHENTIK_REFRESH_TOKEN_URL=""
+
+# Fief
+FIEF_CLIENT_ID=""
+FIEF_CLIENT_SECRET=""
+FIEF_URL=""
+
+# Zulip integration (optional)
+ZULIP_API_KEY=""
+ZULIP_BOT_EMAIL=""
+ZULIP_REALM=""
+
+# External services (optional)
+ZEPHYR_LLM_URL=""
+
+# Redis/KV (optional)
+KV_REST_API_TOKEN=""
+KV_REST_API_READ_ONLY_TOKEN=""
+KV_REST_API_URL=""
+KV_URL=""