docs: Add presence race condition design document

Comprehensive analysis of the race condition where users in the same
Reflector room can end up in different Daily.co rooms.

Contents:
- Problem statement and symptoms
- Evidence from Hypothesis simulation
- Current system analysis with code references
- Detailed race condition timeline
- Why current mitigations (Daily API, fallback) are insufficient
- Three solution options with trade-offs
- Recommended approach: Track "intent to join" via /joining endpoint
- Implementation checklist and file references

Key insight: The race is a data model gap, not a timing issue. Backend
needs explicit knowledge of joining users before Daily presence API
sees them.

.gitleaksignore

@@ -4,4 +4,5 @@ docs/docs/installation/daily-setup.md:curl-auth-header:277
 gpu/self_hosted/DEV_SETUP.md:curl-auth-header:74
 gpu/self_hosted/DEV_SETUP.md:curl-auth-header:83
 server/reflector/worker/process.py:generic-api-key:465
+server/tests/test_recording_request_flow.py:generic-api-key:121
 server/reflector/worker/process.py:generic-api-key:594

CHANGELOG.md

@@ -1,17 +1,5 @@
 # Changelog
 
-## [0.33.0](https://github.com/Monadical-SAS/reflector/compare/v0.32.2...v0.33.0) (2026-02-05)
-
-
-### Features
-
-* Daily+hatchet default ([#846](https://github.com/Monadical-SAS/reflector/issues/846)) ([15ab2e3](https://github.com/Monadical-SAS/reflector/commit/15ab2e306eacf575494b4b5d2b2ad779d44a1c7f))
-
-
-### Bug Fixes
-
-* websocket tests ([#825](https://github.com/Monadical-SAS/reflector/issues/825)) ([1ce1c7a](https://github.com/Monadical-SAS/reflector/commit/1ce1c7a910b6c374115d2437b17f9d288ef094dc))
-
 ## [0.32.2](https://github.com/Monadical-SAS/reflector/compare/v0.32.1...v0.32.2) (2026-02-03)
 
 

COMPOSE_STANDALONE_TODO.md

@@ -1,10 +0,0 @@
-# Standalone Compose: Remaining Production Work
-
-## Server/worker/beat: remove host network mode + bind mounts
-
-Currently `server` uses `network_mode: host` and all three services bind-mount `./server/:/app/`. For full standalone prod:
-
-- Remove `network_mode: host` from server
-- Remove bind-mount volumes from server, worker, beat (use built image only)
-- Update `compose_cmd` in `setup-standalone.sh` to not rely on host network
-- Change `SERVER_API_URL` from `http://host.docker.internal:1250` to `http://server:1250` (server reachable via Docker network once off host mode)

PRESENCE_RACE_DESIGN_DOC.md

@@ -0,0 +1,414 @@
+# Presence System Race Condition: Design Document
+
+## Executive Summary
+
+Users in the same Reflector room can end up in **different Daily.co rooms** due to race conditions in meeting lifecycle management. This document details the root cause, why current mitigations are insufficient, and proposes a solution that eliminates the race by design.
+
+---
+
+## Problem Statement
+
+When a user quickly leaves and rejoins a meeting (e.g., closes tab and reopens within seconds), they may find themselves in a different Daily.co room than other participants in the same Reflector room. This breaks the core assumption that all users in a Reflector room share the same video call.
+
+### Symptoms
+- User A and User B are in the same Reflector room but different Daily.co rooms
+- User reports "I can't see/hear the other participant"
+- Meeting appears active but users are isolated
+
+---
+
+## Evidence: Hypothesis Simulation
+
+A simulation was built to model the presence system and find race conditions through randomized action sequences.
+
+**Location**: `server/tests/simulation/`
+
+```bash
+cd server
+
+# Current system config - finds race conditions
+uv run pytest tests/simulation/test_presence_race.py::test_presence_race_conditions_current_system
+# Result: XFAIL (expected failure - race found)
+
+# Fixed system config - no race conditions
+uv run pytest tests/simulation/test_presence_race.py::test_presence_no_race_conditions_fixed_system
+# Result: PASS
+```
+
+The simulation models:
+- Discrete time clock for deterministic replay
+- Daily.co rooms, participants, presence API with configurable lag
+- Reflector meetings, sessions, webhooks
+- User state machine: `idle → joining → handshaking → connected → leaving → idle`
+- Background tasks: `poll_daily_room_presence`, `process_meetings`
+
+### Key Finding
+
+The simulation proves that **even with the Daily API call**, a race window exists during WebRTC handshake when users are invisible to the presence API.
+
+---
+
+## Current System Analysis
+
+### Architecture Overview
+
+```
+┌─────────────┐     ┌─────────────┐     ┌─────────────┐
+│   Frontend  │────▶│   Backend   │────▶│  Daily.co   │
+│  (Next.js)  │     │  (FastAPI)  │     │    API      │
+└─────────────┘     └─────────────┘     └─────────────┘
+                           │
+                           ▼
+                    ┌─────────────┐
+                    │  Database   │
+                    │  (Sessions) │
+                    └─────────────┘
+```
+
+### Relevant Code Paths
+
+#### 1. Meeting Join Flow
+- **File**: `server/reflector/views/rooms.py`
+- **Endpoint**: `POST /rooms/{room_name}/meeting`
+- Returns existing active meeting or creates new one
+- User then connects to Daily.co via WebRTC (frontend)
+
+#### 2. Presence Polling
+- **File**: `server/reflector/worker/process.py:642`
+- **Function**: `poll_daily_room_presence()`
+- Called by webhooks (`participant.joined`, `participant.left`) and `/joined`, `/leave` endpoints
+- Queries Daily API for current participants
+- Updates `daily_participant_sessions` table in database
+
+#### 3. Meeting Deactivation
+- **File**: `server/reflector/worker/process.py:754`
+- **Function**: `process_meetings()`
+- Runs periodically (every 60s via Celery beat)
+- Checks if meetings should be deactivated
+
+**Current implementation** (lines 806-833):
+```python
+if meeting.platform == "daily":
+    try:
+        presence = await client.get_room_presence(meeting.room_name)
+        has_active_sessions = presence.total_count > 0
+        # ...
+    except Exception:
+        logger_.warning("Daily.co presence API failed, falling back to DB sessions")
+        room_sessions = await client.get_room_sessions(meeting.room_name)
+        has_active_sessions = bool(
+            room_sessions and any(s.ended_at is None for s in room_sessions)
+        )
+```
+
+**Key observation**: The code already uses the Daily API (`get_room_presence`), not just the database. The race condition persists despite this.
+
+### Endpoints from feature-leave-endpoint Branch
+
+The `feature-leave-endpoint` branch added explicit leave/join notifications:
+
+| Endpoint | Purpose | Trigger |
+|----------|---------|---------|
+| `POST /rooms/{room_name}/meetings/{meeting_id}/join` | Get meeting info | User navigates to room |
+| `POST /rooms/{room_name}/meetings/{meeting_id}/joined` | Signal connection complete | After WebRTC connects |
+| `POST /rooms/{room_name}/meetings/{meeting_id}/leave` | Signal user leaving | Tab close via sendBeacon |
+
+These endpoints trigger `poll_daily_room_presence_task` to update session state faster than waiting for webhooks.
+
+---
+
+## Race Condition: Detailed Analysis
+
+### The Fundamental Problem
+
+**The backend has no knowledge of users who are in the process of joining (WebRTC handshake phase).**
+
+Data sources available to backend:
+| Source | What it knows | Limitation |
+|--------|---------------|------------|
+| Daily Presence API | Currently connected users | 0-500ms lag; doesn't see handshaking users |
+| Database sessions | Historical join/leave events | Stale; updated by polls |
+| Webhooks | Join/leave events | Delayed; can fail |
+
+**Gap**: No source knows about users between "decided to join" and "WebRTC handshake complete".
+
+### Race Scenario Timeline
+
+```
+T+0ms:    User A connected to Meeting M1, visible in Daily presence
+T+1000ms: User A closes browser tab
+T+1050ms: participant.left webhook fires → poll_daily_room_presence queued
+T+1500ms: User A reopens tab (quick rejoin)
+T+1600ms: POST /meeting returns M1 (still active)
+T+1700ms: Frontend starts WebRTC handshake
+T+2000ms: User A in handshake - NOT visible to Daily presence API
+T+2100ms: poll runs → sees 0 participants → marks session as left_at
+T+3000ms: process_meetings runs
+T+3100ms: Daily API returns 0 participants (user still handshaking)
+T+3200ms: has_active_sessions=False, has_had_sessions=True
+T+3300ms: Meeting deactivated, Daily room deleted
+T+4000ms: User A WebRTC completes → Daily room is gone!
+T+5000ms: User B joins same Reflector room → new Meeting M2 created
+
+RESULT: User A orphaned, User B in different Daily room
+```
+
+### Why Current Mitigations Are Insufficient
+
+#### 1. Using Daily API (already implemented)
+The code already calls `get_room_presence()` instead of relying solely on database sessions. **This doesn't help** because the Daily presence API itself doesn't see users during WebRTC handshake (0-500ms consistency lag + handshake duration of 500-3000ms).
+
+#### 2. Fallback to Database
+When Daily API fails, the code falls back to database sessions. This is **worse** because database is even more stale than the API.
+
+#### 3. Leave/Join Endpoints
+The `/joined` and `/leave` endpoints trigger immediate polls, reducing the window but **not eliminating it**. The poll still only sees what Daily presence API reports.
+
+---
+
+## Proposed Solutions
+
+### Option A: Grace Period (Not Recommended)
+
+Add a time-based buffer before deactivation.
+
+```python
+GRACE_PERIOD_SECONDS = 10
+
+if not has_active_sessions and has_had_sessions:
+    recent_activity = await get_recent_activity(meeting_id, within_seconds=GRACE_PERIOD_SECONDS)
+    if recent_activity:
+        continue  # Skip deactivation
+```
+
+**Pros:**
+- Simple to implement
+- Low risk
+
+**Cons:**
+- Arbitrary timeout value (why 10s? why not 5s or 30s?)
+- Feels like a hack ("setTimeout solution")
+- Delays legitimate deactivation
+- Doesn't eliminate race, just makes it less likely
+
+### Option B: Track "Intent to Join" (Recommended)
+
+Add explicit state tracking for users who are in the process of joining.
+
+**New endpoint**: `POST /rooms/{room_name}/meetings/{meeting_id}/joining`
+
+**Flow change**:
+```
+Current:
+1. POST /join → get meeting info
+2. Render Daily iframe (start WebRTC)
+3. POST /joined (after connected)
+
+Proposed:
+1. POST /join → get meeting info
+2. POST /joining → "I'm about to connect" ← NEW (wait for 200 OK)
+3. Render Daily iframe (start WebRTC)
+4. POST /joined (after connected)
+```
+
+**Backend tracking**:
+```python
+# On /joining endpoint
+await pending_joins.create(meeting_id=meeting_id, user_id=user_id, created_at=now())
+
+# In process_meetings
+pending = await pending_joins.get_recent(meeting_id, max_age_seconds=30)
+if pending:
+    logger.info("Meeting has pending joins, skipping deactivation")
+    continue
+
+# On /joined endpoint or timeout
+await pending_joins.delete(meeting_id=meeting_id, user_id=user_id)
+```
+
+**Pros:**
+- Eliminates race by design (backend knows before Daily does)
+- Explicit state machine, not time-based guessing
+- Clear semantics
+
+**Cons:**
+- Adds ~50-200ms latency (one round-trip before iframe renders)
+- Requires frontend changes
+- Needs cleanup mechanism for abandoned joins (user closes tab during handshake)
+
+### Option C: Optimistic Locking with Version
+
+Track meeting "version" that must match for deactivation.
+
+**Concept**: Each join attempt increments a version. Deactivation only proceeds if version hasn't changed since presence check.
+
+**Cons:**
+- Complex to implement correctly
+- Still has edge cases with concurrent joins
+
+---
+
+## Recommended Approach: Option B
+
+**Track "Intent to Join"** is the cleanest solution because it:
+
+1. **Eliminates the race by design** - no timing windows
+2. **Makes state explicit** - joining/connected/leaving are tracked, not inferred
+3. **Aligns with existing patterns** - similar to `/joined` and `/leave` endpoints
+4. **No arbitrary timeouts** - unlike grace period
+
+### Data Model Change
+
+Add tracking for pending joins. Options:
+
+| Storage | Pros | Cons |
+|---------|------|------|
+| Redis key | Fast, auto-expire | Lost on Redis restart |
+| Database table | Persistent, queryable | Slightly slower |
+| In-memory | Fastest | Lost on server restart |
+
+**Recommendation**: Redis with TTL (30s expiry) for simplicity. Pending joins are ephemeral - if Redis restarts, worst case is a brief deactivation delay.
+
+```python
+# Redis key format
+pending_join:{meeting_id}:{user_id} = {timestamp}
+# TTL: 30 seconds
+```
+
+### Implementation Checklist
+
+1. **Backend: Add `/joining` endpoint**
+   - File: `server/reflector/views/rooms.py`
+   - Creates Redis key with 30s TTL
+   - Returns 200 OK
+
+2. **Backend: Modify `process_meetings()`**
+   - File: `server/reflector/worker/process.py`
+   - Before deactivation, check for pending joins
+   - If any exist, skip deactivation
+
+3. **Backend: Modify `/joined` endpoint**
+   - Clear pending join on successful connection
+
+4. **Frontend: Call `/joining` before WebRTC**
+   - File: `www/app/[roomName]/components/DailyRoom.tsx`
+   - Await response before rendering Daily iframe
+
+5. **Update simulation**
+   - Add `joining` state tracking to match new design
+   - Verify race condition is eliminated
+
+6. **Integration tests**
+   - Test quick rejoin scenario
+   - Test abandoned join (user closes during handshake)
+   - Test concurrent joins from multiple users
+
+---
+
+## Files Reference
+
+### Core Files to Modify
+| File | Purpose |
+|------|---------|
+| `server/reflector/views/rooms.py` | Add `/joining` endpoint |
+| `server/reflector/worker/process.py` | Check pending joins before deactivation |
+| `www/app/[roomName]/components/DailyRoom.tsx` | Call `/joining` before WebRTC |
+
+### Reference Files
+| File | Contains |
+|------|----------|
+| `server/reflector/video_platforms/daily.py:128` | `get_room_presence()` - Daily API call |
+| `server/reflector/worker/process.py:642` | `poll_daily_room_presence()` - presence polling |
+| `server/reflector/views/daily.py:125` | Webhook handlers |
+| `server/tests/simulation/` | Hypothesis simulation proving the race |
+| `server/tests/test_daily_presence_deactivation.py` | Unit tests for presence logic |
+
+### Simulation Files
+| File | Purpose |
+|------|---------|
+| `tests/simulation/system.py` | Main simulation engine |
+| `tests/simulation/config.py` | Current vs fixed system configs |
+| `tests/simulation/state.py` | State dataclasses |
+| `tests/simulation/test_presence_race.py` | Hypothesis stateful tests |
+| `tests/simulation/test_targeted_scenarios.py` | Specific race scenarios |
+| `server/reflector/presence/model.py` | Shared state machine model |
+
+---
+
+## Alternative Considered: Remove DB Fallback
+
+One simpler change discussed: remove the database fallback when Daily API fails, and "fail loudly" instead.
+
+```python
+# Current (with fallback)
+try:
+    presence = await client.get_room_presence(meeting.room_name)
+    has_active_sessions = presence.total_count > 0
+except Exception:
+    # Fallback to stale DB
+    room_sessions = await client.get_room_sessions(meeting.room_name)
+    has_active_sessions = bool(room_sessions and any(s.ended_at is None for s in room_sessions))
+
+# Proposed (fail loudly)
+try:
+    presence = await client.get_room_presence(meeting.room_name)
+    has_active_sessions = presence.total_count > 0
+except Exception:
+    logger.error("Daily API failed, skipping deactivation check for this meeting")
+    continue  # Don't deactivate if we can't verify
+```
+
+**This helps but doesn't eliminate the race** - it only removes one failure mode (stale DB). The core race (handshake invisibility) remains.
+
+---
+
+## Conclusion
+
+The presence system race condition is a **data model gap**, not a timing issue that can be solved with grace periods. The backend needs explicit knowledge of users who intend to join, before they become visible to the Daily presence API.
+
+The recommended fix is to add a `/joining` endpoint that the frontend calls before starting WebRTC. This creates a "reservation" that prevents premature meeting deactivation during the handshake window.
+
+This approach:
+- Eliminates the race by design
+- Adds minimal latency (~50-200ms)
+- Follows explicit state machine principles
+- Avoids arbitrary timeout hacks
+
+---
+
+## Appendix: Simulation Test Results
+
+```
+$ uv run pytest tests/simulation/ -v
+
+tests/simulation/test_model_conformance.py::TestModelConformance::test_simulation_uses_model_states PASSED
+tests/simulation/test_model_conformance.py::TestModelConformance::test_simulation_respects_transitions PASSED
+tests/simulation/test_model_conformance.py::TestModelConformance::test_simulation_invalid_transitions_checked PASSED
+tests/simulation/test_model_conformance.py::TestModelConformance::test_simulation_implements_protocols PASSED
+tests/simulation/test_model_conformance.py::TestModelConformance::test_simulation_uses_shared_invariants PASSED
+tests/simulation/test_model_conformance.py::TestProductionStateMachine::test_state_machine_has_all_states PASSED
+tests/simulation/test_model_conformance.py::TestProductionStateMachine::test_state_machine_valid_transitions PASSED
+tests/simulation/test_model_conformance.py::TestProductionStateMachine::test_state_machine_invalid_transitions_raise PASSED
+tests/simulation/test_model_conformance.py::TestProductionStateMachine::test_guarded_user_state_transitions PASSED
+tests/simulation/test_model_conformance.py::TestProductionStateMachine::test_guarded_user_state_rejects_invalid PASSED
+tests/simulation/test_model_conformance.py::TestProductionStateMachine::test_guarded_user_state_tracks_history PASSED
+tests/simulation/test_model_conformance.py::TestInvariantConsistency::test_invariants_same_between_model_and_simulation PASSED
+tests/simulation/test_model_conformance.py::test_quick_conformance_check PASSED
+tests/simulation/test_presence_race.py::TestPresenceRaceFixed::runTest PASSED
+tests/simulation/test_presence_race.py::test_presence_race_conditions_current_system XFAIL
+tests/simulation/test_presence_race.py::test_presence_no_race_conditions_fixed_system PASSED
+tests/simulation/test_presence_race.py::test_smoke_presence_simulation PASSED
+tests/simulation/test_targeted_scenarios.py::TestQuickRejoinRace::test_quick_rejoin_causes_split PASSED
+tests/simulation/test_targeted_scenarios.py::TestQuickRejoinRace::test_quick_rejoin_fixed_system PASSED
+tests/simulation/test_targeted_scenarios.py::TestSimultaneousJoins::test_two_users_join_simultaneously PASSED
+tests/simulation/test_targeted_scenarios.py::TestProcessMeetingsRace::test_process_meetings_during_handshake PASSED
+tests/simulation/test_targeted_scenarios.py::TestPresenceLagRace::test_presence_lag_causes_incorrect_count PASSED
+tests/simulation/test_targeted_scenarios.py::TestMeetingDeactivationEdgeCases::test_deactivation_with_no_sessions PASSED
+tests/simulation/test_targeted_scenarios.py::TestMeetingDeactivationEdgeCases::test_deactivation_requires_had_sessions PASSED
+tests/simulation/test_targeted_scenarios.py::TestEventLogTracing::test_event_log_captures_flow PASSED
+tests/simulation/test_targeted_scenarios.py::test_config_presets PASSED
+
+================== 25 passed, 1 xfailed ==================
+```
+
+The `xfail` test (`test_presence_race_conditions_current_system`) demonstrates that the current system configuration has race conditions that can be found through randomized testing.

docker-compose.standalone.yml

@@ -1,129 +0,0 @@
-# Standalone services for fully local deployment (no external dependencies).
-# Usage: docker compose -f docker-compose.yml -f docker-compose.standalone.yml up -d
-#
-# On Linux with NVIDIA GPU, also pass: --profile ollama-gpu
-# On Linux without GPU:                --profile ollama-cpu
-# On Mac: Ollama runs natively (Metal GPU) — no profile needed, services here unused.
-
-services:
-  garage:
-    image: dxflrs/garage:v1.1.0
-    ports:
-      - "3900:3900"   # S3 API
-      - "3903:3903"   # Admin API
-    volumes:
-      - garage_data:/var/lib/garage/data
-      - garage_meta:/var/lib/garage/meta
-      - ./data/garage.toml:/etc/garage.toml:ro
-    restart: unless-stopped
-    healthcheck:
-      test: ["CMD", "/garage", "stats"]
-      interval: 10s
-      timeout: 5s
-      retries: 5
-      start_period: 5s
-
-  ollama:
-    image: ollama/ollama:latest
-    profiles: ["ollama-gpu"]
-    ports:
-      - "11434:11434"
-    volumes:
-      - ollama_data:/root/.ollama
-    deploy:
-      resources:
-        reservations:
-          devices:
-            - driver: nvidia
-              count: all
-              capabilities: [gpu]
-    restart: unless-stopped
-    healthcheck:
-      test: ["CMD", "curl", "-f", "http://localhost:11434/api/tags"]
-      interval: 10s
-      timeout: 5s
-      retries: 5
-
-  ollama-cpu:
-    image: ollama/ollama:latest
-    profiles: ["ollama-cpu"]
-    ports:
-      - "11434:11434"
-    volumes:
-      - ollama_data:/root/.ollama
-    restart: unless-stopped
-    healthcheck:
-      test: ["CMD", "curl", "-f", "http://localhost:11434/api/tags"]
-      interval: 10s
-      timeout: 5s
-      retries: 5
-
-  # Override server/worker/beat to use self-hosted GPU service for transcription+diarization.
-  # compose `environment:` overrides values from `env_file:` — no need to edit server/.env.
-  server:
-    environment:
-      TRANSCRIPT_BACKEND: modal
-      TRANSCRIPT_URL: http://localhost:8100
-      TRANSCRIPT_MODAL_API_KEY: local
-      DIARIZATION_BACKEND: modal
-      DIARIZATION_URL: http://localhost:8100
-
-  worker:
-    environment:
-      TRANSCRIPT_BACKEND: modal
-      TRANSCRIPT_URL: http://cpu:8000
-      TRANSCRIPT_MODAL_API_KEY: local
-      DIARIZATION_BACKEND: modal
-      DIARIZATION_URL: http://cpu:8000
-
-  web:
-    image: reflector-frontend-standalone
-    build:
-      context: ./www
-    command: ["node", "server.js"]
-    volumes: !reset []
-    environment:
-      NODE_ENV: production
-
-  cpu:
-    build:
-      context: ./gpu/self_hosted
-      dockerfile: Dockerfile.cpu
-    ports:
-      - "8100:8000"
-    volumes:
-      - gpu_cache:/root/.cache
-    restart: unless-stopped
-    healthcheck:
-      test: ["CMD", "curl", "-f", "http://localhost:8000/docs"]
-      interval: 15s
-      timeout: 5s
-      retries: 10
-      start_period: 120s
-
-  gpu-nvidia:
-    build:
-      context: ./gpu/self_hosted
-    profiles: ["gpu-nvidia"]
-    volumes:
-      - gpu_cache:/root/.cache
-    deploy:
-      resources:
-        reservations:
-          devices:
-            - driver: nvidia
-              count: all
-              capabilities: [gpu]
-    restart: unless-stopped
-    healthcheck:
-      test: ["CMD", "curl", "-f", "http://localhost:8000/docs"]
-      interval: 15s
-      timeout: 5s
-      retries: 10
-      start_period: 120s
-
-volumes:
-  garage_data:
-  garage_meta:
-  ollama_data:
-  gpu_cache:

docker-compose.yml

@@ -2,7 +2,8 @@ services:
   server:
     build:
       context: server
-    network_mode: host
+    ports:
+      - 1250:1250
     volumes:
       - ./server/:/app/
       - /app/.venv
@@ -10,12 +11,6 @@ services:
       - ./server/.env
     environment:
       ENTRYPOINT: server
-      DATABASE_URL: postgresql+asyncpg://reflector:reflector@localhost:5432/reflector
-      REDIS_HOST: localhost
-      CELERY_BROKER_URL: redis://localhost:6379/1
-      CELERY_RESULT_BACKEND: redis://localhost:6379/1
-      HATCHET_CLIENT_SERVER_URL: http://localhost:8889
-      HATCHET_CLIENT_HOST_PORT: localhost:7078
 
   worker:
     build:
@@ -27,11 +22,6 @@ services:
       - ./server/.env
     environment:
       ENTRYPOINT: worker
-      HATCHET_CLIENT_SERVER_URL: http://hatchet:8888
-      HATCHET_CLIENT_HOST_PORT: hatchet:7077
-    depends_on:
-      redis:
-        condition: service_started
 
   beat:
     build:
@@ -43,9 +33,6 @@ services:
       - ./server/.env
     environment:
       ENTRYPOINT: beat
-    depends_on:
-      redis:
-        condition: service_started
 
   hatchet-worker-cpu:
     build:
@@ -57,8 +44,6 @@ services:
       - ./server/.env
     environment:
       ENTRYPOINT: hatchet-worker-cpu
-      HATCHET_CLIENT_SERVER_URL: http://hatchet:8888
-      HATCHET_CLIENT_HOST_PORT: hatchet:7077
     depends_on:
       hatchet:
         condition: service_healthy
@@ -72,8 +57,6 @@ services:
       - ./server/.env
     environment:
       ENTRYPOINT: hatchet-worker-llm
-      HATCHET_CLIENT_SERVER_URL: http://hatchet:8888
-      HATCHET_CLIENT_HOST_PORT: hatchet:7077
     depends_on:
       hatchet:
         condition: service_healthy
@@ -92,16 +75,10 @@ services:
     volumes:
       - ./www:/app/
       - /app/node_modules
-      - next_cache:/app/.next
     env_file:
       - ./www/.env.local
     environment:
       - NODE_ENV=development
-      - SERVER_API_URL=http://host.docker.internal:1250
-    extra_hosts:
-      - "host.docker.internal:host-gateway"
-    depends_on:
-      - server
 
   postgres:
     image: postgres:17
@@ -117,14 +94,13 @@ services:
       - ./server/docker/init-hatchet-db.sql:/docker-entrypoint-initdb.d/init-hatchet-db.sql:ro
     healthcheck:
       test: ["CMD-SHELL", "pg_isready -d reflector -U reflector"]
-      interval: 5s
-      timeout: 5s
-      retries: 10
-      start_period: 15s
+      interval: 10s
+      timeout: 10s
+      retries: 5
+      start_period: 10s
 
   hatchet:
     image: ghcr.io/hatchet-dev/hatchet/hatchet-lite:latest
-    restart: on-failure
     ports:
       - "8889:8888"
       - "7078:7077"
@@ -132,7 +108,7 @@ services:
       postgres:
         condition: service_healthy
     environment:
-      DATABASE_URL: "postgresql://reflector:reflector@postgres:5432/hatchet?sslmode=disable&connect_timeout=30"
+      DATABASE_URL: "postgresql://reflector:reflector@postgres:5432/hatchet?sslmode=disable"
       SERVER_AUTH_COOKIE_DOMAIN: localhost
       SERVER_AUTH_COOKIE_INSECURE: "t"
       SERVER_GRPC_BIND_ADDRESS: "0.0.0.0"
@@ -152,5 +128,6 @@ services:
       retries: 5
       start_period: 30s
 
-volumes:
-  next_cache:
+networks:
+  default:
+    attachable: true

docs/docs/installation/setup-standalone.md

@@ -1,214 +0,0 @@
----
-sidebar_position: 2
-title: Standalone Local Setup
----
-
-# Standalone Local Setup
-
-**The goal**: a clueless user clones the repo, runs one script, and has a working Reflector instance locally. No cloud accounts, no API keys, no manual env file editing.
-
-```bash
-git clone https://github.com/monadical-sas/reflector.git
-cd reflector
-./scripts/setup-standalone.sh
-```
-
-The script is idempotent — safe to re-run at any time. It detects what's already set up and skips completed steps.
-
-## Prerequisites
-
-- Docker / OrbStack / Docker Desktop (any)
-- Mac (Apple Silicon) or Linux
-- 16GB+ RAM (32GB recommended for 14B LLM models)
-- **Mac only**: [Ollama](https://ollama.com/download) installed (`brew install ollama`)
-
-## What the script does
-
-### 1. LLM inference via Ollama
-
-**Mac**: starts Ollama natively (Metal GPU acceleration). Pulls the LLM model. Docker containers reach it via `host.docker.internal:11434`.
-
-**Linux**: starts containerized Ollama via `docker-compose.standalone.yml` profile (`ollama-gpu` with NVIDIA, `ollama-cpu` without). Pulls model inside the container.
-
-### 2. Environment files
-
-Generates `server/.env` and `www/.env.local` with standalone defaults:
-
-**`server/.env`** — key settings:
-
-| Variable | Value | Why |
-|----------|-------|-----|
-| `DATABASE_URL` | `postgresql+asyncpg://...@postgres:5432/reflector` | Docker-internal hostname |
-| `REDIS_HOST` | `redis` | Docker-internal hostname |
-| `CELERY_BROKER_URL` | `redis://redis:6379/1` | Docker-internal hostname |
-| `AUTH_BACKEND` | `none` | No Authentik in standalone |
-| `TRANSCRIPT_BACKEND` | `modal` | HTTP API to self-hosted CPU service |
-| `TRANSCRIPT_URL` | `http://cpu:8000` | Docker-internal CPU service |
-| `DIARIZATION_BACKEND` | `modal` | HTTP API to self-hosted CPU service |
-| `DIARIZATION_URL` | `http://cpu:8000` | Docker-internal CPU service |
-| `TRANSLATION_BACKEND` | `passthrough` | No Modal |
-| `LLM_URL` | `http://host.docker.internal:11434/v1` (Mac) | Ollama endpoint |
-
-**`www/.env.local`** — key settings:
-
-| Variable | Value |
-|----------|-------|
-| `API_URL` | `http://localhost:1250` |
-| `SERVER_API_URL` | `http://server:1250` |
-| `WEBSOCKET_URL` | `ws://localhost:1250` |
-| `FEATURE_REQUIRE_LOGIN` | `false` |
-| `NEXTAUTH_SECRET` | `standalone-dev-secret-not-for-production` |
-
-If env files already exist (including symlinks from worktree setup), the script resolves symlinks and ensures all standalone-critical vars are set. Existing vars not related to standalone are preserved.
-
-### 3. Object storage (Garage)
-
-Standalone uses [Garage](https://garagehq.deuxfleurs.fr/) — a lightweight S3-compatible object store running in Docker. The setup script starts Garage, initializes the layout, creates a bucket and access key, and writes the credentials to `server/.env`.
-
-**`server/.env`** — storage settings added by the script:
-
-| Variable | Value | Why |
-|----------|-------|-----|
-| `TRANSCRIPT_STORAGE_BACKEND` | `aws` | Uses the S3-compatible storage driver |
-| `TRANSCRIPT_STORAGE_AWS_ENDPOINT_URL` | `http://garage:3900` | Docker-internal Garage S3 API |
-| `TRANSCRIPT_STORAGE_AWS_BUCKET_NAME` | `reflector-media` | Created by the script |
-| `TRANSCRIPT_STORAGE_AWS_REGION` | `garage` | Must match Garage config |
-| `TRANSCRIPT_STORAGE_AWS_ACCESS_KEY_ID` | *(auto-generated)* | Created by `garage key create` |
-| `TRANSCRIPT_STORAGE_AWS_SECRET_ACCESS_KEY` | *(auto-generated)* | Created by `garage key create` |
-
-The `TRANSCRIPT_STORAGE_AWS_ENDPOINT_URL` setting enables S3-compatible backends. When set, the storage driver uses path-style addressing and routes all requests to the custom endpoint. When unset (production AWS), behavior is unchanged.
-
-Garage config template lives at `scripts/garage.toml`. The setup script generates `data/garage.toml` (gitignored) with a random RPC secret and mounts it read-only into the container. Single-node, `replication_factor=1`.
-
-> **Note**: Presigned URLs embed the Garage Docker hostname (`http://garage:3900`). This is fine — the server proxies S3 responses to the browser. Modal GPU workers cannot reach internal Garage, but standalone doesn't use Modal.
-
-### 4. Transcription and diarization
-
-Standalone runs the self-hosted ML service (`gpu/self_hosted/`) in a CPU-only Docker container named `cpu`. This is the same FastAPI service used for Modal.com GPU deployments, but built with `Dockerfile.cpu` (no NVIDIA CUDA dependencies). The compose service is named `cpu` (not `gpu`) to make clear it runs without GPU acceleration; the source code lives in `gpu/self_hosted/` because it's shared with the GPU deployment.
-
-The `modal` backend name is reused — it just means "HTTP API client". Setting `TRANSCRIPT_URL` / `DIARIZATION_URL` to `http://cpu:8000` routes requests to the local container instead of Modal.com.
-
-On first start, the service downloads pyannote speaker diarization models (~1GB) from a public S3 bundle. Models are cached in a Docker volume (`gpu_cache`) so subsequent starts are fast. No HuggingFace token or API key needed.
-
-> **Performance**: CPU-only transcription and diarization work but are slow (~15 min for a 3 min file). For faster processing on Linux with NVIDIA GPU, use `--profile gpu-nvidia` instead (see `docker-compose.standalone.yml`).
-
-### 5. Docker services
-
-```bash
-docker compose up -d postgres redis garage cpu server worker beat web
-```
-
-All services start in a single command. Garage and `cpu` are already started by earlier steps but included for idempotency. No Hatchet in standalone mode — LLM processing (summaries, topics, titles) runs via Celery tasks.
-
-### 6. Database migrations
-
-Run automatically by the `server` container on startup (`runserver.sh` calls `alembic upgrade head`). No manual step needed.
-
-### 7. Health check
-
-Verifies:
-- CPU service responds (transcription + diarization ready)
-- Server responds at `http://localhost:1250/health`
-- Frontend serves at `http://localhost:3000`
-- LLM endpoint reachable from inside containers
-
-## Services
-
-| Service | Port | Purpose |
-|---------|------|---------|
-| `server` | 1250 | FastAPI backend (runs migrations on start) |
-| `web` | 3000 | Next.js frontend |
-| `postgres` | 5432 | PostgreSQL database |
-| `redis` | 6379 | Cache + Celery broker |
-| `garage` | 3900, 3903 | S3-compatible object storage (S3 API + admin API) |
-| `cpu` | — | Self-hosted transcription + diarization (CPU-only) |
-| `worker` | — | Celery worker (live pipeline post-processing) |
-| `beat` | — | Celery beat (scheduled tasks) |
-
-## Testing programmatically
-
-After the setup script completes, verify the full pipeline (upload, transcription, diarization, LLM summary) via the API:
-
-```bash
-# 1. Create a transcript
-TRANSCRIPT_ID=$(curl -s -X POST 'http://localhost:1250/v1/transcripts' \
-  -H 'Content-Type: application/json' \
-  -d '{"name":"test-upload"}' | python3 -c "import sys,json; print(json.load(sys.stdin)['id'])")
-echo "Created: $TRANSCRIPT_ID"
-
-# 2. Upload an audio file (single-chunk upload)
-curl -s "http://localhost:1250/v1/transcripts/${TRANSCRIPT_ID}/record/upload?chunk_number=0&total_chunks=1" \
-  -X POST -F "chunk=@/path/to/audio.mp3"
-
-# 3. Poll until processing completes (status: ended or error)
-while true; do
-  STATUS=$(curl -s "http://localhost:1250/v1/transcripts/${TRANSCRIPT_ID}" \
-    | python3 -c "import sys,json; print(json.load(sys.stdin)['status'])")
-  echo "Status: $STATUS"
-  case "$STATUS" in ended|error) break;; esac
-  sleep 10
-done
-
-# 4. Check the result
-curl -s "http://localhost:1250/v1/transcripts/${TRANSCRIPT_ID}" | python3 -m json.tool
-```
-
-Expected result: status `ended`, auto-generated `title`, `short_summary`, `long_summary`, and `transcript` text with `Speaker 0` / `Speaker 1` labels.
-
-CPU-only processing is slow (~15 min for a 3 min audio file). Diarization finishes in ~3 min, transcription takes the rest.
-
-## Troubleshooting
-
-### Port conflicts (most common issue)
-
-If the frontend or backend behaves unexpectedly (e.g., env vars seem ignored, changes don't take effect), **check for port conflicts first**:
-
-```bash
-# Check what's listening on key ports
-lsof -i :3000   # frontend
-lsof -i :1250   # backend
-lsof -i :5432   # postgres
-lsof -i :3900   # Garage S3 API
-lsof -i :6379   # Redis
-
-# Kill stale processes on a port
-lsof -ti :3000 | xargs kill
-```
-
-Common causes:
-- A stale `next dev` or `pnpm dev` process from another terminal/worktree
-- Another Docker Compose project (different worktree) with containers on the same ports — the setup script only manages its own project; containers from other projects must be stopped manually (`docker ps` to find them, `docker stop` to kill them)
-
-The setup script checks ports 3000, 1250, 5432, 6379, 3900, 3903 for conflicts before starting services. It ignores OrbStack/Docker Desktop port forwarding processes (which always bind these ports but are not real conflicts).
-
-### OrbStack false port-conflict warnings (Mac)
-
-If you use OrbStack as your Docker runtime, `lsof` will show OrbStack binding ports like 3000, 1250, etc. even when no containers are running. This is OrbStack's port forwarding mechanism — not a real conflict. The setup script filters these out automatically.
-
-### Re-enabling authentication
-
-Standalone runs without authentication (`FEATURE_REQUIRE_LOGIN=false`, `AUTH_BACKEND=none`). To re-enable:
-
-1. In `www/.env.local`: set `FEATURE_REQUIRE_LOGIN=true`, uncomment `AUTHENTIK_ISSUER` and `AUTHENTIK_REFRESH_TOKEN_URL`
-2. In `server/.env`: set `AUTH_BACKEND=authentik` (or your backend), configure `AUTH_JWT_AUDIENCE`
-3. Restart: `docker compose -f docker-compose.yml -f docker-compose.standalone.yml up -d --force-recreate web server`
-
-## What's NOT covered
-
-These require external accounts and infrastructure that can't be scripted:
-
-- **Live meeting rooms** — requires Daily.co account, S3 bucket, IAM roles
-- **Authentication** — requires Authentik deployment and OAuth configuration
-- **Hatchet workflows** — requires separate Hatchet setup for multitrack processing
-- **Production deployment** — see [Deployment Guide](./overview)
-
-## Current status
-
-All steps implemented. The setup script handles everything end-to-end:
-
-- Step 1 (Ollama/LLM) — implemented
-- Step 2 (environment files) — implemented
-- Step 3 (object storage / Garage) — implemented
-- Step 4 (transcription/diarization) — implemented (self-hosted GPU service)
-- Steps 5-7 (Docker, migrations, health) — implemented
-- **Unified script**: `scripts/setup-standalone.sh`

gpu/self_hosted/Dockerfile.cpu

@@ -1,39 +0,0 @@
-FROM python:3.12-slim
-
-ENV PYTHONUNBUFFERED=1 \
-    UV_LINK_MODE=copy \
-    UV_NO_CACHE=1
-
-WORKDIR /tmp
-RUN --mount=type=cache,target=/var/cache/apt,sharing=locked \
- --mount=type=cache,target=/var/lib/apt,sharing=locked \
- apt-get update \
- && apt-get install -y \
-    ffmpeg \
-    curl \
-    ca-certificates \
-    gnupg \
-    wget
-ADD https://astral.sh/uv/install.sh /uv-installer.sh
-RUN sh /uv-installer.sh && rm /uv-installer.sh
-ENV PATH="/root/.local/bin/:$PATH"
-
-RUN mkdir -p /app
-WORKDIR /app
-COPY pyproject.toml uv.lock /app/
-
-
-COPY ./app /app/app
-COPY ./main.py /app/
-COPY ./runserver.sh /app/
-
-# prevent uv failing with too many open files on big cpus
-ENV UV_CONCURRENT_INSTALLS=16
-
-# first install
-RUN --mount=type=cache,target=/root/.cache/uv \
-    uv sync --compile-bytecode --locked
-
-EXPOSE 8000
-
-CMD ["sh", "/app/runserver.sh"]

gpu/self_hosted/app/auth.py

@@ -3,14 +3,14 @@ import os
 from fastapi import Depends, HTTPException, status
 from fastapi.security import OAuth2PasswordBearer
 
-oauth2_scheme = OAuth2PasswordBearer(tokenUrl="token", auto_error=False)
+oauth2_scheme = OAuth2PasswordBearer(tokenUrl="token")
 
 
-def apikey_auth(apikey: str | None = Depends(oauth2_scheme)):
+def apikey_auth(apikey: str = Depends(oauth2_scheme)):
     required_key = os.environ.get("REFLECTOR_GPU_APIKEY")
     if not required_key:
         return
-    if apikey and apikey == required_key:
+    if apikey == required_key:
         return
     raise HTTPException(
         status_code=status.HTTP_401_UNAUTHORIZED,

gpu/self_hosted/app/services/diarizer.py

@@ -1,65 +1,10 @@
-import logging
 import os
-import tarfile
 import threading
-from pathlib import Path
-from urllib.request import urlopen
 
 import torch
 import torchaudio
-import yaml
 from pyannote.audio import Pipeline
 
-logger = logging.getLogger(__name__)
-
-S3_BUNDLE_URL = "https://reflector-public.s3.us-east-1.amazonaws.com/pyannote-speaker-diarization-3.1.tar.gz"
-BUNDLE_CACHE_DIR = Path("/root/.cache/pyannote-bundle")
-
-
-def _ensure_model(cache_dir: Path) -> str:
-    """Download and extract S3 model bundle if not cached."""
-    model_dir = cache_dir / "pyannote-speaker-diarization-3.1"
-    config_path = model_dir / "config.yaml"
-
-    if config_path.exists():
-        logger.info("Using cached model bundle at %s", model_dir)
-        return str(model_dir)
-
-    cache_dir.mkdir(parents=True, exist_ok=True)
-    tarball_path = cache_dir / "model.tar.gz"
-
-    logger.info("Downloading model bundle from %s", S3_BUNDLE_URL)
-    with urlopen(S3_BUNDLE_URL) as response, open(tarball_path, "wb") as f:
-        while chunk := response.read(8192):
-            f.write(chunk)
-
-    logger.info("Extracting model bundle")
-    with tarfile.open(tarball_path, "r:gz") as tar:
-        tar.extractall(path=cache_dir, filter="data")
-    tarball_path.unlink()
-
-    _patch_config(model_dir, cache_dir)
-    return str(model_dir)
-
-
-def _patch_config(model_dir: Path, cache_dir: Path) -> None:
-    """Rewrite config.yaml to reference local pytorch_model.bin paths."""
-    config_path = model_dir / "config.yaml"
-    with open(config_path) as f:
-        config = yaml.safe_load(f)
-
-    config["pipeline"]["params"]["segmentation"] = str(
-        cache_dir / "pyannote-segmentation-3.0" / "pytorch_model.bin"
-    )
-    config["pipeline"]["params"]["embedding"] = str(
-        cache_dir / "pyannote-wespeaker-voxceleb-resnet34-LM" / "pytorch_model.bin"
-    )
-
-    with open(config_path, "w") as f:
-        yaml.dump(config, f)
-
-    logger.info("Patched config.yaml with local model paths")
-
 
 class PyannoteDiarizationService:
     def __init__(self):
@@ -69,20 +14,10 @@ class PyannoteDiarizationService:
 
     def load(self):
         self._device = "cuda" if torch.cuda.is_available() else "cpu"
-        hf_token = os.environ.get("HF_TOKEN")
-
-        if hf_token:
-            logger.info("Loading pyannote model from HuggingFace (HF_TOKEN set)")
-            self._pipeline = Pipeline.from_pretrained(
-                "pyannote/speaker-diarization-3.1",
-                use_auth_token=hf_token,
-            )
-        else:
-            logger.info("HF_TOKEN not set — loading model from S3 bundle")
-            model_path = _ensure_model(BUNDLE_CACHE_DIR)
-            config_path = Path(model_path) / "config.yaml"
-            self._pipeline = Pipeline.from_pretrained(str(config_path))
-
+        self._pipeline = Pipeline.from_pretrained(
+            "pyannote/speaker-diarization-3.1",
+            use_auth_token=os.environ.get("HF_TOKEN"),
+        )
         self._pipeline.to(torch.device(self._device))
 
     def diarize_file(self, file_path: str, timestamp: float = 0.0) -> dict:

scripts/garage.toml

@@ -1,14 +0,0 @@
-metadata_dir = "/var/lib/garage/meta"
-data_dir = "/var/lib/garage/data"
-replication_factor = 1
-
-rpc_secret = "__GARAGE_RPC_SECRET__"
-rpc_bind_addr = "[::]:3901"
-
-[s3_api]
-api_bind_addr = "[::]:3900"
-s3_region = "garage"
-root_domain = ".s3.garage.localhost"
-
-[admin]
-api_bind_addr = "[::]:3903"

scripts/setup-standalone.sh

@@ -1,417 +0,0 @@
-#!/usr/bin/env bash
-#
-# Standalone local development setup for Reflector.
-# Takes a fresh clone to a working instance — no cloud accounts, no API keys.
-#
-# Usage:
-#   ./scripts/setup-standalone.sh
-#
-# Idempotent — safe to re-run at any time.
-#
-set -euo pipefail
-
-SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
-ROOT_DIR="$(cd "$SCRIPT_DIR/.." && pwd)"
-
-SERVER_ENV="$ROOT_DIR/server/.env"
-WWW_ENV="$ROOT_DIR/www/.env.local"
-
-MODEL="${LLM_MODEL:-qwen2.5:14b}"
-OLLAMA_PORT="${OLLAMA_PORT:-11434}"
-
-OS="$(uname -s)"
-
-# --- Colors ---
-RED='\033[0;31m'
-GREEN='\033[0;32m'
-YELLOW='\033[1;33m'
-CYAN='\033[0;36m'
-NC='\033[0m'
-
-info()  { echo -e "${CYAN}==>${NC} $*"; }
-ok()    { echo -e "${GREEN}  ✓${NC} $*"; }
-warn()  { echo -e "${YELLOW}  !${NC} $*"; }
-err()   { echo -e "${RED}  ✗${NC} $*" >&2; }
-
-# --- Helpers ---
-
-wait_for_url() {
-    local url="$1" label="$2" retries="${3:-30}" interval="${4:-2}"
-    for i in $(seq 1 "$retries"); do
-        if curl -sf "$url" > /dev/null 2>&1; then
-            return 0
-        fi
-        echo -ne "\r  Waiting for $label... ($i/$retries)"
-        sleep "$interval"
-    done
-    echo ""
-    err "$label not responding at $url after $retries attempts"
-    return 1
-}
-
-env_has_key() {
-    local file="$1" key="$2"
-    grep -q "^${key}=" "$file" 2>/dev/null
-}
-
-env_set() {
-    local file="$1" key="$2" value="$3"
-    if env_has_key "$file" "$key"; then
-        # Replace existing value (portable sed)
-        if [[ "$OS" == "Darwin" ]]; then
-            sed -i '' "s|^${key}=.*|${key}=${value}|" "$file"
-        else
-            sed -i "s|^${key}=.*|${key}=${value}|" "$file"
-        fi
-    else
-        echo "${key}=${value}" >> "$file"
-    fi
-}
-
-resolve_symlink() {
-    local file="$1"
-    if [[ -L "$file" ]]; then
-        warn "$(basename "$file") is a symlink — creating standalone copy"
-        cp -L "$file" "$file.tmp"
-        rm "$file"
-        mv "$file.tmp" "$file"
-    fi
-}
-
-compose_cmd() {
-    local compose_files="-f $ROOT_DIR/docker-compose.yml -f $ROOT_DIR/docker-compose.standalone.yml"
-    if [[ "$OS" == "Linux" ]] && [[ -n "${OLLAMA_PROFILE:-}" ]]; then
-        docker compose $compose_files --profile "$OLLAMA_PROFILE" "$@"
-    else
-        docker compose $compose_files "$@"
-    fi
-}
-
-# =========================================================
-# Step 1: LLM / Ollama
-# =========================================================
-step_llm() {
-    info "Step 1: LLM setup (Ollama + $MODEL)"
-
-    case "$OS" in
-        Darwin)
-            if ! command -v ollama &> /dev/null; then
-                err "Ollama not found. Install it:"
-                err "  brew install ollama"
-                err "  # or https://ollama.com/download"
-                exit 1
-            fi
-
-            # Start if not running
-            if ! curl -sf "http://localhost:$OLLAMA_PORT/api/tags" > /dev/null 2>&1; then
-                info "Starting Ollama..."
-                ollama serve &
-                disown
-            fi
-
-            wait_for_url "http://localhost:$OLLAMA_PORT/api/tags" "Ollama"
-            echo ""
-
-            # Pull model if not already present
-            if ollama list 2>/dev/null | awk '{print $1}' | grep -qx "$MODEL"; then
-                ok "Model $MODEL already pulled"
-            else
-                info "Pulling model $MODEL (this may take a while)..."
-                ollama pull "$MODEL"
-            fi
-
-            LLM_URL_VALUE="http://host.docker.internal:$OLLAMA_PORT/v1"
-            ;;
-
-        Linux)
-            if command -v nvidia-smi &> /dev/null && nvidia-smi > /dev/null 2>&1; then
-                ok "NVIDIA GPU detected — using ollama-gpu profile"
-                OLLAMA_PROFILE="ollama-gpu"
-                OLLAMA_SVC="ollama"
-                LLM_URL_VALUE="http://ollama:$OLLAMA_PORT/v1"
-            else
-                warn "No NVIDIA GPU — using ollama-cpu profile"
-                OLLAMA_PROFILE="ollama-cpu"
-                OLLAMA_SVC="ollama-cpu"
-                LLM_URL_VALUE="http://ollama-cpu:$OLLAMA_PORT/v1"
-            fi
-
-            info "Starting Ollama container..."
-            compose_cmd up -d
-
-            wait_for_url "http://localhost:$OLLAMA_PORT/api/tags" "Ollama"
-            echo ""
-
-            # Pull model inside container
-            if compose_cmd exec "$OLLAMA_SVC" ollama list 2>/dev/null | awk '{print $1}' | grep -qx "$MODEL"; then
-                ok "Model $MODEL already pulled"
-            else
-                info "Pulling model $MODEL inside container (this may take a while)..."
-                compose_cmd exec "$OLLAMA_SVC" ollama pull "$MODEL"
-            fi
-            ;;
-
-        *)
-            err "Unsupported OS: $OS"
-            exit 1
-            ;;
-    esac
-
-    ok "LLM ready ($MODEL via Ollama)"
-}
-
-# =========================================================
-# Step 2: Generate server/.env
-# =========================================================
-step_server_env() {
-    info "Step 2: Generating server/.env"
-
-    resolve_symlink "$SERVER_ENV"
-
-    if [[ -f "$SERVER_ENV" ]]; then
-        ok "server/.env already exists — ensuring standalone vars"
-    else
-        cat > "$SERVER_ENV" << 'ENVEOF'
-# Generated by setup-standalone.sh — standalone local development
-# Source of truth for settings: server/reflector/settings.py
-ENVEOF
-        ok "Created server/.env"
-    fi
-
-    # Ensure all standalone-critical vars (appends if missing, replaces if present)
-    env_set "$SERVER_ENV" "DATABASE_URL" "postgresql+asyncpg://reflector:reflector@postgres:5432/reflector"
-    env_set "$SERVER_ENV" "REDIS_HOST" "redis"
-    env_set "$SERVER_ENV" "CELERY_BROKER_URL" "redis://redis:6379/1"
-    env_set "$SERVER_ENV" "CELERY_RESULT_BACKEND" "redis://redis:6379/1"
-    env_set "$SERVER_ENV" "AUTH_BACKEND" "none"
-    env_set "$SERVER_ENV" "PUBLIC_MODE" "true"
-    # TRANSCRIPT_BACKEND, TRANSCRIPT_URL, DIARIZATION_BACKEND, DIARIZATION_URL
-    # are set via docker-compose.standalone.yml `environment:` overrides — not written here
-    # so we don't clobber the user's server/.env for non-standalone use.
-    env_set "$SERVER_ENV" "TRANSLATION_BACKEND" "passthrough"
-    env_set "$SERVER_ENV" "LLM_URL" "$LLM_URL_VALUE"
-    env_set "$SERVER_ENV" "LLM_MODEL" "$MODEL"
-    env_set "$SERVER_ENV" "LLM_API_KEY" "not-needed"
-
-    ok "Standalone vars set (LLM_URL=$LLM_URL_VALUE)"
-}
-
-# =========================================================
-# Step 3: Object storage (Garage)
-# =========================================================
-step_storage() {
-    info "Step 3: Object storage (Garage)"
-
-    # Generate garage.toml from template (fill in RPC secret)
-    GARAGE_TOML="$ROOT_DIR/scripts/garage.toml"
-    GARAGE_TOML_RUNTIME="$ROOT_DIR/data/garage.toml"
-    if [[ ! -f "$GARAGE_TOML_RUNTIME" ]]; then
-        mkdir -p "$ROOT_DIR/data"
-        RPC_SECRET=$(openssl rand -hex 32)
-        sed "s|__GARAGE_RPC_SECRET__|${RPC_SECRET}|" "$GARAGE_TOML" > "$GARAGE_TOML_RUNTIME"
-    fi
-
-    compose_cmd up -d garage
-
-    wait_for_url "http://localhost:3903/health" "Garage admin API"
-    echo ""
-
-    # Layout: get node ID, assign, apply (skip if already applied)
-    NODE_ID=$(compose_cmd exec -T garage /garage node id -q 2>/dev/null | tr -d '[:space:]')
-    LAYOUT_STATUS=$(compose_cmd exec -T garage /garage layout show 2>&1 || true)
-    if echo "$LAYOUT_STATUS" | grep -q "No nodes"; then
-        compose_cmd exec -T garage /garage layout assign "$NODE_ID" -c 1G -z dc1
-        compose_cmd exec -T garage /garage layout apply --version 1
-    fi
-
-    # Create bucket (idempotent — skip if exists)
-    if ! compose_cmd exec -T garage /garage bucket info reflector-media &>/dev/null; then
-        compose_cmd exec -T garage /garage bucket create reflector-media
-    fi
-
-    # Create key (idempotent — skip if exists)
-    CREATED_KEY=false
-    if compose_cmd exec -T garage /garage key info reflector &>/dev/null; then
-        ok "Key 'reflector' already exists"
-    else
-        KEY_OUTPUT=$(compose_cmd exec -T garage /garage key create reflector)
-        CREATED_KEY=true
-    fi
-
-    # Grant bucket permissions (idempotent)
-    compose_cmd exec -T garage /garage bucket allow reflector-media --read --write --key reflector
-
-    # Set env vars (only parse key on first create — key info redacts the secret)
-    env_set "$SERVER_ENV" "TRANSCRIPT_STORAGE_BACKEND" "aws"
-    env_set "$SERVER_ENV" "TRANSCRIPT_STORAGE_AWS_ENDPOINT_URL" "http://garage:3900"
-    env_set "$SERVER_ENV" "TRANSCRIPT_STORAGE_AWS_BUCKET_NAME" "reflector-media"
-    env_set "$SERVER_ENV" "TRANSCRIPT_STORAGE_AWS_REGION" "garage"
-    if [[ "$CREATED_KEY" == "true" ]]; then
-        KEY_ID=$(echo "$KEY_OUTPUT" | grep -i "key id" | awk '{print $NF}')
-        KEY_SECRET=$(echo "$KEY_OUTPUT" | grep -i "secret key" | awk '{print $NF}')
-        env_set "$SERVER_ENV" "TRANSCRIPT_STORAGE_AWS_ACCESS_KEY_ID" "$KEY_ID"
-        env_set "$SERVER_ENV" "TRANSCRIPT_STORAGE_AWS_SECRET_ACCESS_KEY" "$KEY_SECRET"
-    fi
-
-    ok "Object storage ready (Garage)"
-}
-
-# =========================================================
-# Step 4: Generate www/.env.local
-# =========================================================
-step_www_env() {
-    info "Step 4: Generating www/.env.local"
-
-    resolve_symlink "$WWW_ENV"
-
-    if [[ -f "$WWW_ENV" ]]; then
-        ok "www/.env.local already exists — ensuring standalone vars"
-    else
-        cat > "$WWW_ENV" << 'ENVEOF'
-# Generated by setup-standalone.sh — standalone local development
-ENVEOF
-        ok "Created www/.env.local"
-    fi
-
-    env_set "$WWW_ENV" "SITE_URL" "http://localhost:3000"
-    env_set "$WWW_ENV" "NEXTAUTH_URL" "http://localhost:3000"
-    env_set "$WWW_ENV" "NEXTAUTH_SECRET" "standalone-dev-secret-not-for-production"
-    env_set "$WWW_ENV" "API_URL" "http://localhost:1250"
-    env_set "$WWW_ENV" "WEBSOCKET_URL" "ws://localhost:1250"
-    env_set "$WWW_ENV" "SERVER_API_URL" "http://server:1250"
-    env_set "$WWW_ENV" "FEATURE_REQUIRE_LOGIN" "false"
-
-    ok "Standalone www vars set"
-}
-
-# =========================================================
-# Step 5: Start all services
-# =========================================================
-step_services() {
-    info "Step 5: Starting Docker services"
-
-    # Check for port conflicts — stale processes silently shadow Docker port mappings.
-    # OrbStack/Docker Desktop bind ports for forwarding; ignore those PIDs.
-    local ports_ok=true
-    for port in 3000 1250 5432 6379 3900 3903; do
-        local pids
-        pids=$(lsof -ti :"$port" 2>/dev/null || true)
-        for pid in $pids; do
-            local pname
-            pname=$(ps -p "$pid" -o comm= 2>/dev/null || true)
-            # OrbStack and Docker Desktop own port forwarding — not real conflicts
-            if [[ "$pname" == *"OrbStack"* ]] || [[ "$pname" == *"com.docker"* ]] || [[ "$pname" == *"vpnkit"* ]]; then
-                continue
-            fi
-            warn "Port $port already in use by PID $pid ($pname)"
-            warn "Kill it with: lsof -ti :$port | xargs kill"
-            ports_ok=false
-        done
-    done
-    if [[ "$ports_ok" == "false" ]]; then
-        warn "Port conflicts detected — Docker containers may not be reachable"
-        warn "Continuing anyway (services will start but may be shadowed)"
-    fi
-
-    # server runs alembic migrations on startup automatically (see runserver.sh)
-    compose_cmd up -d postgres redis garage cpu server worker beat web
-    ok "Containers started"
-    info "Server is running migrations (alembic upgrade head)..."
-}
-
-# =========================================================
-# Step 6: Health checks
-# =========================================================
-step_health() {
-    info "Step 6: Health checks"
-
-    # CPU service may take a while on first start (model download + load).
-    # No host port exposed — check via docker exec.
-    info "Waiting for CPU service (first start downloads ~1GB of models)..."
-    local cpu_ok=false
-    for i in $(seq 1 120); do
-        if compose_cmd exec -T cpu curl -sf http://localhost:8000/docs > /dev/null 2>&1; then
-            cpu_ok=true
-            break
-        fi
-        echo -ne "\r  Waiting for CPU service... ($i/120)"
-        sleep 5
-    done
-    echo ""
-    if [[ "$cpu_ok" == "true" ]]; then
-        ok "CPU service healthy (transcription + diarization)"
-    else
-        warn "CPU service not ready yet — it will keep loading in the background"
-        warn "Check with: docker compose logs cpu"
-    fi
-
-    wait_for_url "http://localhost:1250/health" "Server API" 60 3
-    echo ""
-    ok "Server API healthy"
-
-    wait_for_url "http://localhost:3000" "Frontend" 90 3
-    echo ""
-    ok "Frontend responding"
-
-    # Check LLM reachability from inside a container
-    if compose_cmd exec -T server \
-        curl -sf "$LLM_URL_VALUE/models" > /dev/null 2>&1; then
-        ok "LLM reachable from containers"
-    else
-        warn "LLM not reachable from containers at $LLM_URL_VALUE"
-        warn "Summaries/topics/titles won't work until LLM is accessible"
-    fi
-}
-
-# =========================================================
-# Main
-# =========================================================
-main() {
-    echo ""
-    echo "=========================================="
-    echo " Reflector — Standalone Local Setup"
-    echo "=========================================="
-    echo ""
-
-    # Ensure we're in the repo root
-    if [[ ! -f "$ROOT_DIR/docker-compose.yml" ]]; then
-        err "docker-compose.yml not found in $ROOT_DIR"
-        err "Run this script from the repo root: ./scripts/setup-standalone.sh"
-        exit 1
-    fi
-
-
-    # LLM_URL_VALUE is set by step_llm, used by later steps
-    LLM_URL_VALUE=""
-    OLLAMA_PROFILE=""
-
-    # docker-compose.yml may reference env_files that don't exist yet;
-    # touch them so compose_cmd works before the steps that populate them.
-    touch "$SERVER_ENV" "$WWW_ENV"
-
-    step_llm
-    echo ""
-    step_server_env
-    echo ""
-    step_storage
-    echo ""
-    step_www_env
-    echo ""
-    step_services
-    echo ""
-    step_health
-
-    echo ""
-    echo "=========================================="
-    echo -e " ${GREEN}Reflector is running!${NC}"
-    echo "=========================================="
-    echo ""
-    echo "  Frontend:  http://localhost:3000"
-    echo "  API:       http://localhost:1250"
-    echo ""
-    echo "  To stop:   docker compose down"
-    echo "  To re-run: ./scripts/setup-standalone.sh"
-    echo ""
-}
-
-main "$@"

server/.env.example

@@ -66,22 +66,15 @@ TRANSLATE_URL=https://monadical-sas--reflector-translator-web.modal.run
 ## LLM backend (Required)
 ##
 ## Responsible for generating titles, summaries, and topic detection
-## Supports any OpenAI-compatible endpoint.
+## Requires OpenAI API key
 ## =======================================================
 
-## --- Option A: Local LLM via Ollama (recommended for dev) ---
-## Setup: ./scripts/setup-standalone.sh
-## Mac: Ollama runs natively (Metal GPU). Containers reach it via host.docker.internal.
-## Linux: docker compose --profile ollama-gpu up -d (or ollama-cpu for no GPU)
-LLM_URL=http://host.docker.internal:11434/v1
-LLM_MODEL=qwen2.5:14b
-LLM_API_KEY=not-needed
-## Linux with containerized Ollama: LLM_URL=http://ollama:11434/v1
+## OpenAI API key - get from https://platform.openai.com/account/api-keys
+LLM_API_KEY=sk-your-openai-api-key
+LLM_MODEL=gpt-4o-mini
 
-## --- Option B: Remote/cloud LLM ---
-#LLM_API_KEY=sk-your-openai-api-key
-#LLM_MODEL=gpt-4o-mini
-## LLM_URL defaults to OpenAI when unset
+## Optional: Custom endpoint (defaults to OpenAI)
+# LLM_URL=https://api.openai.com/v1
 
 ## Context size for summary generation (tokens)
 LLM_CONTEXT_WINDOW=16000

server/docs/DAILY_REFLECTOR_DATA_MODEL.md

@@ -86,7 +86,7 @@ Daily.co Room: "daily-private-igor-20260110042117"
 | **Purpose** | Tracks active session state | Links recordings, transcripts, participants |
 | **Scope** | Per room instance | Per Reflector room + timestamp |
 
-**Critical Limitation:** Daily.co's recordings API often does NOT return `mtgSessionId`, requiring time-based matching (see [Time-Based Matching](#time-based-matching)).
+**Critical Limitation:** Daily.co's recordings API often does NOT return `mtgSessionId` (can be null), requiring time-based matching (see [Time-Based Matching](#time-based-matching)).
 
 ### Recording
 
@@ -101,6 +101,30 @@ Daily.co Room: "daily-private-igor-20260110042117"
 
 **Critical Behavior:** Recording **stops/restarts** create **separate recording objects** with unique IDs.
 
+### instanceId (Reflector-Generated)
+
+**Definition:** UUID we generate and send when starting recording via REST API.
+
+**Generation:** Deterministic from meeting_id
+- Cloud: `instanceId = meeting_id` directly
+- Raw-tracks: `instanceId = UUIDv5(meeting_id, namespace)`
+
+**Key behaviors:**
+- ✅ **Reuse allowed:** Same instanceId can be used after stop (validated 2026-01-20)
+- ❌ **Not returned:** Daily.co does NOT echo instanceId back in GET /recordings response
+- ✅ **Present in error webhooks:** `recording.error` webhook includes instanceId
+- **Purpose:** Allows multiple concurrent recordings (cloud + raw-tracks) in same room
+
+**Stop/restart example:**
+```
+Recording 1: POST /start with instanceId="779e6376..." → recording_id="ee00c4e8..."
+Stop recording
+Recording 2: POST /start with instanceId="779e6376..." (SAME) → recording_id="b702f509..." (DIFFERENT)
+✅ Both succeed, different recording_ids returned
+```
+
+**Implication:** Cannot match recordings by instanceId (not in response) - must use recording_id.
+
 ---
 
 ## Entity Relationships
@@ -196,6 +220,19 @@ Daily.co Room: "daily-private-igor-20260110042117"
 
 `mtgSessionId` identifies a **Daily.co meeting session** (not individual participants, not a room).
 
+**Reliability:** Can be null or present in GET /recordings response (unreliable).
+
+**When present:** Multiple recordings from same session (stop/restart with participants connected) share same mtgSessionId.
+
+**Example (validated 2026-01-20):**
+```json
+Recording 1: {"id": "ee00c4e8...", "mtgSessionId": "92c4136a-a8da-41c5-9c45-e9a2baae6bd6"}
+Recording 2: {"id": "b702f509...", "mtgSessionId": "92c4136a-a8da-41c5-9c45-e9a2baae6bd6"}
+// Same mtgSessionId (stop/restart in same session)
+```
+
+**When null:** Common - Daily.co API does not reliably populate this field.
+
 ### session_id (Per-Participant)
 
 **Different concept:** Per-participant connection identifier from webhooks.
@@ -220,16 +257,24 @@ TABLE daily_participant_session (
 
 Daily.co's recordings API does not reliably return `mtgSessionId`, making it impossible to directly link recordings to meetings via Daily.co's identifiers.
 
-**Example API response:**
+**Example API response (mtgSessionId can be null OR present):**
 ```json
 {
   "id": "recording-uuid",
   "room_name": "daily-private-igor-20260110042117",
   "start_ts": 1768018896,
-  "mtgSessionId": null  ← Missing!
+  "mtgSessionId": null  // ← Often null (unreliable)
+}
+
+// OR (when present):
+{
+  "id": "recording-uuid",
+  "mtgSessionId": "92c4136a-a8da-41c5-9c45-e9a2baae6bd6"  // ← Sometimes present
 }
 ```
 
+**Key insight:** Cannot rely on mtgSessionId for matching (unreliable). instanceId also not returned. Only reliable identifier is recording.id.
+
 ### Solution: Time-Based Matching
 
 **Implementation:** `reflector/db/meetings.py:get_by_room_name_and_time()`
@@ -491,6 +536,10 @@ UI: User sees 3 separate transcripts
 
 
 ---
-**Document Version:** 1.0
-**Last Verified:** 2026-01-15
-**Data Source:** Production database + Daily.co API inspection
+**Document Version:** 1.1
+**Last Updated:** 2026-01-20
+**Data Source:** Production database + Daily.co API inspection + empirical testing
+**Changes in 1.1:**
+- Added instanceId behavior documentation (reuse allowed, not returned in API)
+- Clarified mtgSessionId reliability (can be null or present)
+- Added empirical validation of stop/restart behavior

server/migrations/versions/3aa20b96d963_drop_use_celery_column.py

@@ -1,35 +0,0 @@
-"""drop_use_celery_column
-
-Revision ID: 3aa20b96d963
-Revises: e69f08ead8ea
-Create Date: 2026-02-05 10:12:44.065279
-
-"""
-
-from typing import Sequence, Union
-
-import sqlalchemy as sa
-from alembic import op
-
-# revision identifiers, used by Alembic.
-revision: str = "3aa20b96d963"
-down_revision: Union[str, None] = "e69f08ead8ea"
-branch_labels: Union[str, Sequence[str], None] = None
-depends_on: Union[str, Sequence[str], None] = None
-
-
-def upgrade() -> None:
-    with op.batch_alter_table("room", schema=None) as batch_op:
-        batch_op.drop_column("use_celery")
-
-
-def downgrade() -> None:
-    with op.batch_alter_table("room", schema=None) as batch_op:
-        batch_op.add_column(
-            sa.Column(
-                "use_celery",
-                sa.Boolean(),
-                server_default=sa.text("false"),
-                nullable=False,
-            )
-        )

server/migrations/versions/f5b008fa8a14_add_daily_recording_requests.py

@@ -0,0 +1,67 @@
+"""add_daily_recording_requests
+
+Revision ID: f5b008fa8a14
+Revises: 1b1e6a6fc465
+Create Date: 2026-01-20 22:32:06.697144
+
+"""
+
+from typing import Sequence, Union
+
+import sqlalchemy as sa
+from alembic import op
+
+# revision identifiers, used by Alembic.
+revision: str = "f5b008fa8a14"
+down_revision: Union[str, None] = "1b1e6a6fc465"
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+
+def upgrade() -> None:
+    op.create_table(
+        "daily_recording_request",
+        sa.Column("recording_id", sa.String(), nullable=False),
+        sa.Column("meeting_id", sa.String(), nullable=False),
+        sa.Column("instance_id", sa.String(), nullable=False),
+        sa.Column("type", sa.String(), nullable=False),
+        sa.Column("requested_at", sa.DateTime(timezone=True), nullable=False),
+        sa.ForeignKeyConstraint(["meeting_id"], ["meeting.id"], ondelete="CASCADE"),
+        sa.PrimaryKeyConstraint("recording_id"),
+    )
+    op.create_index("idx_meeting_id", "daily_recording_request", ["meeting_id"])
+    op.create_index("idx_instance_id", "daily_recording_request", ["instance_id"])
+
+    # Clean up orphaned recordings before adding FK constraint
+    op.execute("""
+        UPDATE recording SET status = 'orphan', meeting_id = NULL
+        WHERE meeting_id IS NOT NULL
+        AND meeting_id NOT IN (SELECT id FROM meeting)
+    """)
+
+    # Add FK constraint to recording table (cascade delete recordings when meeting deleted)
+    op.execute("""
+        ALTER TABLE recording ADD CONSTRAINT fk_recording_meeting
+          FOREIGN KEY (meeting_id) REFERENCES meeting(id) ON DELETE CASCADE
+    """)
+
+    # Add CHECK constraints to enforce orphan invariants
+    op.execute("""
+        ALTER TABLE recording ADD CONSTRAINT chk_orphan_no_meeting
+          CHECK (status != 'orphan' OR meeting_id IS NULL)
+    """)
+    op.execute("""
+        ALTER TABLE recording ADD CONSTRAINT chk_non_orphan_has_meeting
+          CHECK (status = 'orphan' OR meeting_id IS NOT NULL)
+    """)
+
+
+def downgrade() -> None:
+    op.execute("ALTER TABLE recording DROP CONSTRAINT IF EXISTS chk_orphan_no_meeting")
+    op.execute(
+        "ALTER TABLE recording DROP CONSTRAINT IF EXISTS chk_non_orphan_has_meeting"
+    )
+    op.execute("ALTER TABLE recording DROP CONSTRAINT IF EXISTS fk_recording_meeting")
+    op.drop_index("idx_instance_id", table_name="daily_recording_request")
+    op.drop_index("idx_meeting_id", table_name="daily_recording_request")
+    op.drop_table("daily_recording_request")

server/pyproject.toml

@@ -68,6 +68,7 @@ evaluation = [
     "pydantic>=2.1.1",
 ]
 local = [
+    "pyannote-audio>=3.3.2",
     "faster-whisper>=0.10.0",
 ]
 silero-vad = [

server/reflector/asynctask.py

@@ -22,8 +22,6 @@ def asynctask(f):
                 await database.disconnect()
 
         coro = run_with_db()
-        if current_task:
-            return asyncio.run(coro)
         try:
             loop = asyncio.get_running_loop()
         except RuntimeError:

server/reflector/auth/auth_none.py

@@ -1,5 +1,11 @@
+from typing import Annotated
+
+from fastapi import Depends
+from fastapi.security import OAuth2PasswordBearer
 from pydantic import BaseModel
 
+oauth2_scheme = OAuth2PasswordBearer(tokenUrl="token", auto_error=False)
+
 
 class UserInfo(BaseModel):
     sub: str
@@ -9,13 +15,13 @@ class AccessTokenInfo(BaseModel):
     pass
 
 
-def authenticated():
+def authenticated(token: Annotated[str, Depends(oauth2_scheme)]):
     return None
 
 
-def current_user():
+def current_user(token: Annotated[str, Depends(oauth2_scheme)]):
     return None
 
 
-def current_user_optional():
+def current_user_optional(token: Annotated[str, Depends(oauth2_scheme)]):
     return None

server/reflector/dailyco_api/client.py

@@ -146,8 +146,6 @@ class DailyApiClient:
             )
             raise DailyApiError(operation, response)
 
-        if not response.content:
-            return {}
         return response.json()
 
     # ============================================================================

server/reflector/dailyco_api/recording_orphans.py

@@ -0,0 +1,56 @@
+"""Utility for creating orphan recordings."""
+
+import os
+from datetime import datetime, timezone
+
+from reflector.db.recordings import Recording, recordings_controller
+from reflector.logger import logger
+from reflector.utils.string import NonEmptyString
+
+
+async def create_and_log_orphan(
+    recording_id: NonEmptyString,
+    bucket_name: str,
+    room_name: str,
+    start_ts: int,
+    track_keys: list[str] | None,
+    source: str,
+) -> bool:
+    """Create orphan recording and log if first occurrence.
+
+    Args:
+        recording_id: Daily.co recording ID
+        bucket_name: S3 bucket (empty string for cloud recordings)
+        room_name: Daily.co room name
+        start_ts: Unix timestamp
+        track_keys: Track keys for raw-tracks, None for cloud
+        source: "webhook" or "polling" for logging
+
+    Returns:
+        True if created (first poller), False if already exists
+    """
+    if track_keys:
+        object_key = os.path.dirname(track_keys[0]) if track_keys else room_name
+    else:
+        object_key = room_name
+
+    created = await recordings_controller.create_orphan(
+        Recording(
+            id=recording_id,
+            bucket_name=bucket_name,
+            object_key=object_key,
+            recorded_at=datetime.fromtimestamp(start_ts, tz=timezone.utc),
+            track_keys=track_keys,
+            meeting_id=None,
+            status="orphan",
+        )
+    )
+
+    if created:
+        logger.error(
+            f"Orphan recording ({source})",
+            recording_id=recording_id,
+            room_name=room_name,
+        )
+
+    return created

server/reflector/dailyco_api/webhook_utils.py

@@ -99,7 +99,7 @@ def extract_room_name(event: DailyWebhookEvent) -> str | None:
         >>> event = DailyWebhookEvent(**webhook_payload)
         >>> room_name = extract_room_name(event)
     """
-    room = event.payload.get("room_name") or event.payload.get("room")
+    room = event.payload.get("room_name")
     # Ensure we return a string, not any falsy value that might be in payload
     return room if isinstance(room, str) else None
 

server/reflector/dailyco_api/webhooks.py

@@ -6,7 +6,7 @@ Reference: https://docs.daily.co/reference/rest-api/webhooks
 
 from typing import Annotated, Any, Dict, Literal, Union
 
-from pydantic import AliasChoices, BaseModel, ConfigDict, Field, field_validator
+from pydantic import BaseModel, Field, field_validator
 
 from reflector.utils.string import NonEmptyString
 
@@ -41,8 +41,6 @@ class DailyTrack(BaseModel):
     Reference: https://docs.daily.co/reference/rest-api/recordings
     """
 
-    model_config = ConfigDict(extra="ignore")
-
     type: Literal["audio", "video"]
     s3Key: NonEmptyString = Field(description="S3 object key for the track file")
     size: int = Field(description="File size in bytes")
@@ -56,8 +54,6 @@ class DailyWebhookEvent(BaseModel):
     Reference: https://docs.daily.co/reference/rest-api/webhooks
     """
 
-    model_config = ConfigDict(extra="ignore")
-
     version: NonEmptyString = Field(
         description="Represents the version of the event. This uses semantic versioning to inform a consumer if the payload has introduced any breaking changes"
     )
@@ -86,13 +82,7 @@ class ParticipantJoinedPayload(BaseModel):
     Reference: https://docs.daily.co/reference/rest-api/webhooks/events/participant-joined
     """
 
-    model_config = ConfigDict(extra="ignore")
-
-    room_name: NonEmptyString | None = Field(
-        None,
-        description="Daily.co room name",
-        validation_alias=AliasChoices("room_name", "room"),
-    )
+    room_name: NonEmptyString | None = Field(None, description="Daily.co room name")
     session_id: NonEmptyString = Field(description="Daily.co session identifier")
     user_id: NonEmptyString = Field(description="User identifier (may be encoded)")
     user_name: NonEmptyString | None = Field(None, description="User display name")
@@ -110,13 +100,7 @@ class ParticipantLeftPayload(BaseModel):
     Reference: https://docs.daily.co/reference/rest-api/webhooks/events/participant-left
     """
 
-    model_config = ConfigDict(extra="ignore")
-
-    room_name: NonEmptyString | None = Field(
-        None,
-        description="Daily.co room name",
-        validation_alias=AliasChoices("room_name", "room"),
-    )
+    room_name: NonEmptyString | None = Field(None, description="Daily.co room name")
     session_id: NonEmptyString = Field(description="Daily.co session identifier")
     user_id: NonEmptyString = Field(description="User identifier (may be encoded)")
     user_name: NonEmptyString | None = Field(None, description="User display name")
@@ -128,9 +112,6 @@ class ParticipantLeftPayload(BaseModel):
     _normalize_joined_at = field_validator("joined_at", mode="before")(
         normalize_timestamp_to_int
     )
-    _normalize_duration = field_validator("duration", mode="before")(
-        normalize_timestamp_to_int
-    )
 
 
 class RecordingStartedPayload(BaseModel):
@@ -140,8 +121,6 @@ class RecordingStartedPayload(BaseModel):
     Reference: https://docs.daily.co/reference/rest-api/webhooks/events/recording-started
     """
 
-    model_config = ConfigDict(extra="ignore")
-
     room_name: NonEmptyString | None = Field(None, description="Daily.co room name")
     recording_id: NonEmptyString = Field(description="Recording identifier")
     start_ts: int | None = Field(None, description="Recording start timestamp")
@@ -159,9 +138,7 @@ class RecordingReadyToDownloadPayload(BaseModel):
     Reference: https://docs.daily.co/reference/rest-api/webhooks/events/recording-ready-to-download
     """
 
-    model_config = ConfigDict(extra="ignore")
-
-    type: Literal["cloud", "cloud-audio-only", "raw-tracks"] = Field(
+    type: Literal["cloud", "raw-tracks"] = Field(
         description="The type of recording that was generated"
     )
     recording_id: NonEmptyString = Field(
@@ -176,9 +153,8 @@ class RecordingReadyToDownloadPayload(BaseModel):
     status: Literal["finished"] = Field(
         description="The status of the given recording (always 'finished' in ready-to-download webhook, see RecordingStatus in responses.py for full API statuses)"
     )
-    max_participants: int | None = Field(
-        None,
-        description="The number of participants on the call that were recorded (optional; Daily may omit it in some webhook versions)",
+    max_participants: int = Field(
+        description="The number of participants on the call that were recorded"
     )
     duration: int = Field(description="The duration in seconds of the call")
     s3_key: NonEmptyString = Field(
@@ -204,8 +180,6 @@ class RecordingErrorPayload(BaseModel):
     Reference: https://docs.daily.co/reference/rest-api/webhooks/events/recording-error
     """
 
-    model_config = ConfigDict(extra="ignore")
-
     action: Literal["clourd-recording-err", "cloud-recording-error"] = Field(
         description="A string describing the event that was emitted (both variants are documented)"
     )
@@ -226,8 +200,6 @@ class RecordingErrorPayload(BaseModel):
 
 
 class ParticipantJoinedEvent(BaseModel):
-    model_config = ConfigDict(extra="ignore")
-
     version: NonEmptyString
     type: Literal["participant.joined"]
     id: NonEmptyString
@@ -240,8 +212,6 @@ class ParticipantJoinedEvent(BaseModel):
 
 
 class ParticipantLeftEvent(BaseModel):
-    model_config = ConfigDict(extra="ignore")
-
     version: NonEmptyString
     type: Literal["participant.left"]
     id: NonEmptyString
@@ -254,8 +224,6 @@ class ParticipantLeftEvent(BaseModel):
 
 
 class RecordingStartedEvent(BaseModel):
-    model_config = ConfigDict(extra="ignore")
-
     version: NonEmptyString
     type: Literal["recording.started"]
     id: NonEmptyString
@@ -268,8 +236,6 @@ class RecordingStartedEvent(BaseModel):
 
 
 class RecordingReadyEvent(BaseModel):
-    model_config = ConfigDict(extra="ignore")
-
     version: NonEmptyString
     type: Literal["recording.ready-to-download"]
     id: NonEmptyString
@@ -282,8 +248,6 @@ class RecordingReadyEvent(BaseModel):
 
 
 class RecordingErrorEvent(BaseModel):
-    model_config = ConfigDict(extra="ignore")
-
     version: NonEmptyString
     type: Literal["recording.error"]
     id: NonEmptyString

server/reflector/db/__init__.py

@@ -26,6 +26,7 @@ def get_database() -> databases.Database:
 # import models
 import reflector.db.calendar_events  # noqa
 import reflector.db.daily_participant_sessions  # noqa
+import reflector.db.daily_recording_requests  # noqa
 import reflector.db.meetings  # noqa
 import reflector.db.recordings  # noqa
 import reflector.db.rooms  # noqa

server/reflector/db/daily_recording_requests.py

@@ -0,0 +1,111 @@
+from datetime import datetime
+from typing import Literal
+from uuid import UUID
+
+import sqlalchemy as sa
+from pydantic import BaseModel
+from sqlalchemy.dialects.postgresql import insert
+
+from reflector.db import get_database, metadata
+from reflector.utils.string import NonEmptyString
+
+daily_recording_requests = sa.Table(
+    "daily_recording_request",
+    metadata,
+    sa.Column("recording_id", sa.String, primary_key=True),
+    sa.Column(
+        "meeting_id",
+        sa.String,
+        sa.ForeignKey("meeting.id", ondelete="CASCADE"),
+        nullable=False,
+    ),
+    sa.Column("instance_id", sa.String, nullable=False),
+    sa.Column("type", sa.String, nullable=False),
+    sa.Column("requested_at", sa.DateTime(timezone=True), nullable=False),
+    sa.Index("idx_meeting_id", "meeting_id"),
+    sa.Index("idx_instance_id", "instance_id"),
+)
+
+
+class DailyRecordingRequest(BaseModel):
+    recording_id: NonEmptyString
+    meeting_id: NonEmptyString
+    instance_id: UUID
+    type: Literal["cloud", "raw-tracks"]
+    requested_at: datetime
+
+
+class DailyRecordingRequestsController:
+    async def create(self, request: DailyRecordingRequest) -> None:
+        stmt = insert(daily_recording_requests).values(
+            recording_id=request.recording_id,
+            meeting_id=request.meeting_id,
+            instance_id=str(request.instance_id),
+            type=request.type,
+            requested_at=request.requested_at,
+        )
+        stmt = stmt.on_conflict_do_nothing(index_elements=["recording_id"])
+        await get_database().execute(stmt)
+
+    async def find_by_recording_id(
+        self,
+        recording_id: NonEmptyString,
+    ) -> tuple[NonEmptyString, Literal["cloud", "raw-tracks"]] | None:
+        query = daily_recording_requests.select().where(
+            daily_recording_requests.c.recording_id == recording_id
+        )
+        result = await get_database().fetch_one(query)
+
+        if not result:
+            return None
+
+        req = DailyRecordingRequest(
+            recording_id=result["recording_id"],
+            meeting_id=result["meeting_id"],
+            instance_id=UUID(result["instance_id"]),
+            type=result["type"],
+            requested_at=result["requested_at"],
+        )
+        return (req.meeting_id, req.type)
+
+    async def find_by_instance_id(
+        self,
+        instance_id: UUID,
+    ) -> list[DailyRecordingRequest]:
+        """Multiple recordings can have same instance_id (stop/restart)."""
+        query = daily_recording_requests.select().where(
+            daily_recording_requests.c.instance_id == str(instance_id)
+        )
+        results = await get_database().fetch_all(query)
+        return [
+            DailyRecordingRequest(
+                recording_id=r["recording_id"],
+                meeting_id=r["meeting_id"],
+                instance_id=UUID(r["instance_id"]),
+                type=r["type"],
+                requested_at=r["requested_at"],
+            )
+            for r in results
+        ]
+
+    async def get_by_meeting_id(
+        self,
+        meeting_id: NonEmptyString,
+    ) -> list[DailyRecordingRequest]:
+        query = daily_recording_requests.select().where(
+            daily_recording_requests.c.meeting_id == meeting_id
+        )
+        results = await get_database().fetch_all(query)
+        return [
+            DailyRecordingRequest(
+                recording_id=r["recording_id"],
+                meeting_id=r["meeting_id"],
+                instance_id=UUID(r["instance_id"]),
+                type=r["type"],
+                requested_at=r["requested_at"],
+            )
+            for r in results
+        ]
+
+
+daily_recording_requests_controller = DailyRecordingRequestsController()

server/reflector/db/meetings.py

@@ -1,4 +1,4 @@
-from datetime import datetime, timedelta
+from datetime import datetime
 from typing import Any, Literal
 
 import sqlalchemy as sa
@@ -183,84 +183,6 @@ class MeetingController:
         results = await get_database().fetch_all(query)
         return [Meeting(**r) for r in results]
 
-    async def get_by_room_name_and_time(
-        self,
-        room_name: NonEmptyString,
-        recording_start: datetime,
-        time_window_hours: int = 168,
-    ) -> Meeting | None:
-        """
-        Get meeting by room name closest to recording timestamp.
-
-        HACK ALERT: Daily.co doesn't return instanceId in recordings API response,
-        and mtgSessionId is separate from our instanceId. Time-based matching is
-        the least-bad workaround.
-
-        This handles edge case of duplicate room_name values in DB (race conditions,
-        double-clicks, etc.) by matching based on temporal proximity.
-
-        Algorithm:
-        1. Find meetings within time_window_hours of recording_start
-        2. Return meeting with start_date closest to recording_start
-        3. If tie, return first by meeting.id (deterministic)
-
-        Args:
-            room_name: Daily.co room name from recording
-            recording_start: Timezone-aware datetime from recording.start_ts
-            time_window_hours: Search window (default 168 = 1 week)
-
-        Returns:
-            Meeting closest to recording timestamp, or None if no matches
-
-        Failure modes:
-        - Multiple meetings in same room within ~5 minutes: picks closest
-        - All meetings outside time window: returns None
-        - Clock skew between Daily.co and DB: 1-week window tolerates this
-
-        Why 1 week window:
-        - Handles webhook failures (recording discovered days later)
-        - Tolerates clock skew
-        - Rejects unrelated meetings from weeks ago
-
-        """
-        # Validate timezone-aware datetime
-        if recording_start.tzinfo is None:
-            raise ValueError(
-                f"recording_start must be timezone-aware, got naive datetime: {recording_start}"
-            )
-
-        window_start = recording_start - timedelta(hours=time_window_hours)
-        window_end = recording_start + timedelta(hours=time_window_hours)
-
-        query = (
-            meetings.select()
-            .where(
-                sa.and_(
-                    meetings.c.room_name == room_name,
-                    meetings.c.start_date >= window_start,
-                    meetings.c.start_date <= window_end,
-                )
-            )
-            .order_by(meetings.c.start_date)
-        )
-
-        results = await get_database().fetch_all(query)
-        if not results:
-            return None
-
-        candidates = [Meeting(**r) for r in results]
-
-        # Find meeting with start_date closest to recording_start
-        closest = min(
-            candidates,
-            key=lambda m: (
-                abs((m.start_date - recording_start).total_seconds()),
-                m.id,  # Tie-breaker: deterministic by UUID
-            ),
-        )
-
-        return closest
-
     async def get_active(self, room: Room, current_time: datetime) -> Meeting | None:
         """
         Get latest active meeting for a room.
@@ -350,44 +272,6 @@ class MeetingController:
         query = meetings.update().where(meetings.c.id == meeting_id).values(**kwargs)
         await get_database().execute(query)
 
-    async def set_cloud_recording_if_missing(
-        self,
-        meeting_id: NonEmptyString,
-        s3_key: NonEmptyString,
-        duration: int,
-    ) -> bool:
-        """
-        Set cloud recording only if not already set.
-
-        Returns True if updated, False if already set.
-        Prevents webhook/polling race condition via atomic WHERE clause.
-        """
-        # Check current value before update to detect actual change
-        meeting_before = await self.get_by_id(meeting_id)
-        if not meeting_before:
-            return False
-
-        was_null = meeting_before.daily_composed_video_s3_key is None
-
-        query = (
-            meetings.update()
-            .where(
-                sa.and_(
-                    meetings.c.id == meeting_id,
-                    meetings.c.daily_composed_video_s3_key.is_(None),
-                )
-            )
-            .values(
-                daily_composed_video_s3_key=s3_key,
-                daily_composed_video_duration=duration,
-            )
-        )
-        await get_database().execute(query)
-
-        # Return True only if value was NULL before (actual update occurred)
-        # If was_null=False, the WHERE clause prevented the update
-        return was_null
-
     async def increment_num_clients(self, meeting_id: str) -> None:
         """Atomically increment participant count."""
         query = (
@@ -467,6 +351,27 @@ class MeetingConsentController:
         result = await get_database().fetch_one(query)
         return result is not None
 
+    async def set_cloud_recording_if_missing(
+        self,
+        meeting_id: NonEmptyString,
+        s3_key: NonEmptyString,
+        duration: int,
+    ) -> bool:
+        """Returns True if updated, False if already set."""
+        query = (
+            meetings.update()
+            .where(
+                meetings.c.id == meeting_id,
+                meetings.c.daily_composed_video_s3_key.is_(None),
+            )
+            .values(
+                daily_composed_video_s3_key=s3_key,
+                daily_composed_video_duration=duration,
+            )
+        )
+        result = await get_database().execute(query)
+        return result.rowcount > 0
+
 
 meetings_controller = MeetingController()
 meeting_consent_controller = MeetingConsentController()

server/reflector/db/recordings.py

@@ -4,10 +4,10 @@ from typing import Literal
 import sqlalchemy as sa
 from pydantic import BaseModel, Field
 from sqlalchemy import or_
+from sqlalchemy.dialects.postgresql import insert
 
 from reflector.db import get_database, metadata
 from reflector.utils import generate_uuid4
-from reflector.utils.string import NonEmptyString
 
 recordings = sa.Table(
     "recording",
@@ -31,14 +31,13 @@ recordings = sa.Table(
 class Recording(BaseModel):
     id: str = Field(default_factory=generate_uuid4)
     bucket_name: str
-    # for single-track
     object_key: str
     recorded_at: datetime
-    status: Literal["pending", "processing", "completed", "failed"] = "pending"
+    status: Literal["pending", "processing", "completed", "failed", "orphan"] = (
+        "pending"
+    )
     meeting_id: str | None = None
-    # for multitrack reprocessing
-    # track_keys can be empty list [] if recording finished but no audio was captured (silence/muted)
-    # None means not a multitrack recording, [] means multitrack with no tracks
+    # None = single-track, [] = multitrack with no audio, [keys...] = multitrack with audio
     track_keys: list[str] | None = None
 
     @property
@@ -72,20 +71,6 @@ class RecordingController:
         query = recordings.delete().where(recordings.c.id == id)
         await get_database().execute(query)
 
-    async def set_meeting_id(
-        self,
-        recording_id: NonEmptyString,
-        meeting_id: NonEmptyString,
-    ) -> None:
-        """Link recording to meeting."""
-        query = (
-            recordings.update()
-            .where(recordings.c.id == recording_id)
-            .values(meeting_id=meeting_id)
-        )
-        await get_database().execute(query)
-
-    # no check for existence
     async def get_by_ids(self, recording_ids: list[str]) -> list[Recording]:
         if not recording_ids:
             return []
@@ -104,9 +89,12 @@ class RecordingController:
 
         This is more efficient than fetching all recordings and filtering in Python.
         """
-        from reflector.db.transcripts import (
-            transcripts,  # noqa: PLC0415 cyclic import
-        )
+        # INLINE IMPORT REQUIRED: Circular dependency
+        # - recordings.py needs transcripts table for JOIN query
+        # - transcripts.py imports recordings_controller
+        # - db/__init__.py loads recordings before transcripts (line 31 vs 33)
+        # - Top-level import would fail during module initialization
+        from reflector.db.transcripts import transcripts
 
         query = (
             recordings.select()
@@ -124,5 +112,27 @@ class RecordingController:
         recordings_list = [Recording(**row) for row in results]
         return [r for r in recordings_list if r.is_multitrack]
 
+    async def try_create_with_meeting(self, recording: Recording) -> bool:
+        """Returns True if created, False if already exists."""
+        assert recording.meeting_id is not None, "meeting_id required for non-orphan"
+        assert recording.status != "orphan", "use create_orphan for orphans"
+
+        stmt = insert(recordings).values(**recording.model_dump())
+        stmt = stmt.on_conflict_do_nothing(index_elements=["id"])
+        result = await get_database().execute(stmt)
+
+        return result.rowcount > 0
+
+    async def create_orphan(self, recording: Recording) -> bool:
+        """Returns True if created, False if already exists."""
+        assert recording.status == "orphan", "status must be 'orphan'"
+        assert recording.meeting_id is None, "meeting_id must be NULL for orphan"
+
+        stmt = insert(recordings).values(**recording.model_dump())
+        stmt = stmt.on_conflict_do_nothing(index_elements=["id"])
+        result = await get_database().execute(stmt)
+
+        return result.rowcount > 0
+
 
 recordings_controller = RecordingController()

server/reflector/db/rooms.py

@@ -57,6 +57,12 @@ rooms = sqlalchemy.Table(
         sqlalchemy.String,
         nullable=False,
     ),
+    sqlalchemy.Column(
+        "use_celery",
+        sqlalchemy.Boolean,
+        nullable=False,
+        server_default=false(),
+    ),
     sqlalchemy.Column(
         "skip_consent",
         sqlalchemy.Boolean,
@@ -91,6 +97,7 @@ class Room(BaseModel):
     ics_last_sync: datetime | None = None
     ics_last_etag: str | None = None
     platform: Platform = Field(default_factory=lambda: settings.DEFAULT_VIDEO_PLATFORM)
+    use_celery: bool = False
     skip_consent: bool = False
 
 

server/reflector/db/search.py

@@ -26,7 +26,6 @@ from reflector.db.rooms import rooms
 from reflector.db.transcripts import SourceKind, TranscriptStatus, transcripts
 from reflector.db.utils import is_postgresql
 from reflector.logger import logger
-from reflector.settings import settings
 from reflector.utils.string import NonEmptyString, try_parse_non_empty_string
 
 DEFAULT_SEARCH_LIMIT = 20
@@ -397,7 +396,7 @@ class SearchController:
                     transcripts.c.user_id == params.user_id, rooms.c.is_shared
                 )
             )
-        elif not settings.PUBLIC_MODE:
+        else:
             base_query = base_query.where(rooms.c.is_shared)
         if params.room_id:
             base_query = base_query.where(transcripts.c.room_id == params.room_id)

server/reflector/db/transcripts.py

@@ -406,7 +406,7 @@ class TranscriptController:
             query = query.where(
                 or_(transcripts.c.user_id == user_id, rooms.c.is_shared)
             )
-        elif not settings.PUBLIC_MODE:
+        else:
             query = query.where(rooms.c.is_shared)
 
         if source_kind:

server/reflector/hatchet/client.py

@@ -12,9 +12,7 @@ import threading
 
 from hatchet_sdk import ClientConfig, Hatchet
 from hatchet_sdk.clients.rest.models import V1TaskStatus
-from hatchet_sdk.rate_limit import RateLimitDuration
 
-from reflector.hatchet.constants import LLM_RATE_LIMIT_KEY, LLM_RATE_LIMIT_PER_SECOND
 from reflector.logger import logger
 from reflector.settings import settings
 
@@ -115,26 +113,3 @@ class HatchetClientManager:
         """Reset the client instance (for testing)."""
         with cls._lock:
             cls._instance = None
-
-    @classmethod
-    async def ensure_rate_limit(cls) -> None:
-        """Ensure the LLM rate limit exists in Hatchet.
-
-        Uses the Hatchet SDK rate_limits client (aio_put). See:
-        https://docs.hatchet.run/sdks/python/feature-clients/rate_limits
-        """
-        logger.info(
-            "[Hatchet] Ensuring rate limit exists",
-            rate_limit_key=LLM_RATE_LIMIT_KEY,
-            limit=LLM_RATE_LIMIT_PER_SECOND,
-        )
-        client = cls.get_client()
-        await client.rate_limits.aio_put(
-            key=LLM_RATE_LIMIT_KEY,
-            limit=LLM_RATE_LIMIT_PER_SECOND,
-            duration=RateLimitDuration.SECOND,
-        )
-        logger.info(
-            "[Hatchet] Rate limit put successfully",
-            rate_limit_key=LLM_RATE_LIMIT_KEY,
-        )

server/reflector/hatchet/run_workers_llm.py

@@ -3,8 +3,6 @@ LLM/I/O worker pool for all non-CPU tasks.
 Handles: all tasks except mixdown_tracks (transcription, LLM inference, orchestration)
 """
 
-import asyncio
-
 from reflector.hatchet.client import HatchetClientManager
 from reflector.hatchet.workflows.daily_multitrack_pipeline import (
     daily_multitrack_pipeline,
@@ -22,15 +20,6 @@ POOL = "llm-io"
 def main():
     hatchet = HatchetClientManager.get_client()
 
-    try:
-        asyncio.run(HatchetClientManager.ensure_rate_limit())
-    except Exception as e:
-        logger.warning(
-            "[Hatchet] Rate limit initialization failed, but continuing. "
-            "If workflows fail to register, rate limits may need to be created manually.",
-            error=str(e),
-        )
-
     logger.info(
         "Starting Hatchet LLM worker pool (all tasks except mixdown)",
         worker_name=WORKER_NAME,

server/reflector/hatchet/workflows/daily_multitrack_pipeline.py

@@ -171,13 +171,11 @@ async def set_workflow_error_status(transcript_id: NonEmptyString) -> bool:
 
 def _spawn_storage():
     """Create fresh storage instance."""
-    # TODO: replace direct AwsStorage construction with get_transcripts_storage() factory
     return AwsStorage(
         aws_bucket_name=settings.TRANSCRIPT_STORAGE_AWS_BUCKET_NAME,
         aws_region=settings.TRANSCRIPT_STORAGE_AWS_REGION,
         aws_access_key_id=settings.TRANSCRIPT_STORAGE_AWS_ACCESS_KEY_ID,
         aws_secret_access_key=settings.TRANSCRIPT_STORAGE_AWS_SECRET_ACCESS_KEY,
-        aws_endpoint_url=settings.TRANSCRIPT_STORAGE_AWS_ENDPOINT_URL,
     )
 
 

server/reflector/hatchet/workflows/padding_workflow.py

@@ -49,13 +49,11 @@ async def pad_track(input: PaddingInput, ctx: Context) -> PadTrackResult:
         from reflector.settings import settings  # noqa: PLC0415
         from reflector.storage.storage_aws import AwsStorage  # noqa: PLC0415
 
-        # TODO: replace direct AwsStorage construction with get_transcripts_storage() factory
         storage = AwsStorage(
             aws_bucket_name=settings.TRANSCRIPT_STORAGE_AWS_BUCKET_NAME,
             aws_region=settings.TRANSCRIPT_STORAGE_AWS_REGION,
             aws_access_key_id=settings.TRANSCRIPT_STORAGE_AWS_ACCESS_KEY_ID,
             aws_secret_access_key=settings.TRANSCRIPT_STORAGE_AWS_SECRET_ACCESS_KEY,
-            aws_endpoint_url=settings.TRANSCRIPT_STORAGE_AWS_ENDPOINT_URL,
         )
 
         source_url = await storage.get_file_url(

server/reflector/hatchet/workflows/track_processing.py

@@ -60,7 +60,6 @@ async def pad_track(input: TrackInput, ctx: Context) -> PadTrackResult:
 
     try:
         # Create fresh storage instance to avoid aioboto3 fork issues
-        # TODO: replace direct AwsStorage construction with get_transcripts_storage() factory
         from reflector.settings import settings  # noqa: PLC0415
         from reflector.storage.storage_aws import AwsStorage  # noqa: PLC0415
 
@@ -69,7 +68,6 @@ async def pad_track(input: TrackInput, ctx: Context) -> PadTrackResult:
             aws_region=settings.TRANSCRIPT_STORAGE_AWS_REGION,
             aws_access_key_id=settings.TRANSCRIPT_STORAGE_AWS_ACCESS_KEY_ID,
             aws_secret_access_key=settings.TRANSCRIPT_STORAGE_AWS_SECRET_ACCESS_KEY,
-            aws_endpoint_url=settings.TRANSCRIPT_STORAGE_AWS_ENDPOINT_URL,
         )
 
         source_url = await storage.get_file_url(
@@ -161,7 +159,6 @@ async def transcribe_track(input: TrackInput, ctx: Context) -> TranscribeTrackRe
             raise ValueError("Missing padded_key from pad_track")
 
         # Presign URL on demand (avoids stale URLs on workflow replay)
-        # TODO: replace direct AwsStorage construction with get_transcripts_storage() factory
         from reflector.settings import settings  # noqa: PLC0415
         from reflector.storage.storage_aws import AwsStorage  # noqa: PLC0415
 
@@ -170,7 +167,6 @@ async def transcribe_track(input: TrackInput, ctx: Context) -> TranscribeTrackRe
             aws_region=settings.TRANSCRIPT_STORAGE_AWS_REGION,
             aws_access_key_id=settings.TRANSCRIPT_STORAGE_AWS_ACCESS_KEY_ID,
             aws_secret_access_key=settings.TRANSCRIPT_STORAGE_AWS_SECRET_ACCESS_KEY,
-            aws_endpoint_url=settings.TRANSCRIPT_STORAGE_AWS_ENDPOINT_URL,
         )
 
         audio_url = await storage.get_file_url(

server/reflector/llm.py

@@ -144,18 +144,7 @@ class StructuredOutputWorkflow(Workflow, Generic[OutputT]):
         )
 
         # Network retries handled by OpenAILike (max_retries=3)
-        # response_format enables grammar-based constrained decoding on backends
-        # that support it (DMR/llama.cpp, vLLM, Ollama, OpenAI).
-        response = await Settings.llm.acomplete(
-            json_prompt,
-            response_format={
-                "type": "json_schema",
-                "json_schema": {
-                    "name": self.output_cls.__name__,
-                    "schema": self.output_cls.model_json_schema(),
-                },
-            },
-        )
+        response = await Settings.llm.acomplete(json_prompt)
         return ExtractionDone(output=response.text)
 
     @step

server/reflector/processors/audio_diarization_pyannote.py

@@ -0,0 +1,74 @@
+import os
+
+import torch
+import torchaudio
+from pyannote.audio import Pipeline
+
+from reflector.processors.audio_diarization import AudioDiarizationProcessor
+from reflector.processors.audio_diarization_auto import AudioDiarizationAutoProcessor
+from reflector.processors.types import AudioDiarizationInput, DiarizationSegment
+
+
+class AudioDiarizationPyannoteProcessor(AudioDiarizationProcessor):
+    """Local diarization processor using pyannote.audio library"""
+
+    def __init__(
+        self,
+        model_name: str = "pyannote/speaker-diarization-3.1",
+        pyannote_auth_token: str | None = None,
+        device: str | None = None,
+        **kwargs,
+    ):
+        super().__init__(**kwargs)
+        self.model_name = model_name
+        self.auth_token = pyannote_auth_token or os.environ.get("HF_TOKEN")
+        self.device = device
+
+        if device is None:
+            self.device = "cuda" if torch.cuda.is_available() else "cpu"
+
+        self.logger.info(f"Loading pyannote diarization model: {self.model_name}")
+        self.diarization_pipeline = Pipeline.from_pretrained(
+            self.model_name, use_auth_token=self.auth_token
+        )
+        self.diarization_pipeline.to(torch.device(self.device))
+        self.logger.info(f"Diarization model loaded on device: {self.device}")
+
+    async def _diarize(self, data: AudioDiarizationInput) -> list[DiarizationSegment]:
+        try:
+            # Load audio file (audio_url is assumed to be a local file path)
+            self.logger.info(f"Loading local audio file: {data.audio_url}")
+            waveform, sample_rate = torchaudio.load(data.audio_url)
+            audio_input = {"waveform": waveform, "sample_rate": sample_rate}
+            self.logger.info("Running speaker diarization")
+            diarization = self.diarization_pipeline(audio_input)
+
+            # Convert pyannote diarization output to our format
+            segments = []
+            for segment, _, speaker in diarization.itertracks(yield_label=True):
+                # Extract speaker number from label (e.g., "SPEAKER_00" -> 0)
+                speaker_id = 0
+                if speaker.startswith("SPEAKER_"):
+                    try:
+                        speaker_id = int(speaker.split("_")[-1])
+                    except (ValueError, IndexError):
+                        # Fallback to hash-based ID if parsing fails
+                        speaker_id = hash(speaker) % 1000
+
+                segments.append(
+                    {
+                        "start": round(segment.start, 3),
+                        "end": round(segment.end, 3),
+                        "speaker": speaker_id,
+                    }
+                )
+
+            self.logger.info(f"Diarization completed with {len(segments)} segments")
+            return segments
+
+        except Exception as e:
+            self.logger.exception(f"Diarization failed: {e}")
+            raise
+
+
+AudioDiarizationAutoProcessor.register("pyannote", AudioDiarizationPyannoteProcessor)

server/reflector/services/transcript_process.py

@@ -15,10 +15,14 @@ from hatchet_sdk.clients.rest.exceptions import ApiException, NotFoundException
 from hatchet_sdk.clients.rest.models import V1TaskStatus
 
 from reflector.db.recordings import recordings_controller
+from reflector.db.rooms import rooms_controller
 from reflector.db.transcripts import Transcript, transcripts_controller
 from reflector.hatchet.client import HatchetClientManager
 from reflector.logger import logger
 from reflector.pipelines.main_file_pipeline import task_pipeline_file_process
+from reflector.pipelines.main_multitrack_pipeline import (
+    task_pipeline_multitrack_process,
+)
 from reflector.utils.string import NonEmptyString
 
 
@@ -97,11 +101,8 @@ async def validate_transcript_for_processing(
     if transcript.locked:
         return ValidationLocked(detail="Recording is locked")
 
-    if (
-        transcript.status == "idle"
-        and not transcript.workflow_run_id
-        and not transcript.recording_id
-    ):
+    # Check if recording is ready for processing
+    if transcript.status == "idle" and not transcript.workflow_run_id:
         return ValidationNotReady(detail="Recording is not ready for processing")
 
     # Check Celery tasks
@@ -180,98 +181,124 @@ async def dispatch_transcript_processing(
     Returns AsyncResult for Celery tasks, None for Hatchet workflows.
     """
     if isinstance(config, MultitrackProcessingConfig):
-        # Multitrack processing always uses Hatchet (no Celery fallback)
-        # First check if we can replay (outside transaction since it's read-only)
-        transcript = await transcripts_controller.get_by_id(config.transcript_id)
-        if transcript and transcript.workflow_run_id and not force:
-            can_replay = await HatchetClientManager.can_replay(
-                transcript.workflow_run_id
+        use_celery = False
+        if config.room_id:
+            room = await rooms_controller.get_by_id(config.room_id)
+            use_celery = room.use_celery if room else False
+
+        use_hatchet = not use_celery
+
+        if use_celery:
+            logger.info(
+                "Room uses legacy Celery processing",
+                room_id=config.room_id,
+                transcript_id=config.transcript_id,
             )
-            if can_replay:
-                await HatchetClientManager.replay_workflow(transcript.workflow_run_id)
-                logger.info(
-                    "Replaying Hatchet workflow",
-                    workflow_id=transcript.workflow_run_id,
+
+        if use_hatchet:
+            # First check if we can replay (outside transaction since it's read-only)
+            transcript = await transcripts_controller.get_by_id(config.transcript_id)
+            if transcript and transcript.workflow_run_id and not force:
+                can_replay = await HatchetClientManager.can_replay(
+                    transcript.workflow_run_id
                 )
-                return None
-            else:
-                # Workflow can't replay (CANCELLED, COMPLETED, or 404 deleted)
-                # Log and proceed to start new workflow
+                if can_replay:
+                    await HatchetClientManager.replay_workflow(
+                        transcript.workflow_run_id
+                    )
+                    logger.info(
+                        "Replaying Hatchet workflow",
+                        workflow_id=transcript.workflow_run_id,
+                    )
+                    return None
+                else:
+                    # Workflow can't replay (CANCELLED, COMPLETED, or 404 deleted)
+                    # Log and proceed to start new workflow
+                    try:
+                        status = await HatchetClientManager.get_workflow_run_status(
+                            transcript.workflow_run_id
+                        )
+                        logger.info(
+                            "Old workflow not replayable, starting new",
+                            old_workflow_id=transcript.workflow_run_id,
+                            old_status=status.value,
+                        )
+                    except NotFoundException:
+                        # Workflow deleted from Hatchet but ID still in DB
+                        logger.info(
+                            "Old workflow not found in Hatchet, starting new",
+                            old_workflow_id=transcript.workflow_run_id,
+                        )
+
+            # Force: cancel old workflow if exists
+            if force and transcript and transcript.workflow_run_id:
+                try:
+                    await HatchetClientManager.cancel_workflow(
+                        transcript.workflow_run_id
+                    )
+                    logger.info(
+                        "Cancelled old workflow (--force)",
+                        workflow_id=transcript.workflow_run_id,
+                    )
+                except NotFoundException:
+                    logger.info(
+                        "Old workflow already deleted (--force)",
+                        workflow_id=transcript.workflow_run_id,
+                    )
+                await transcripts_controller.update(
+                    transcript, {"workflow_run_id": None}
+                )
+
+            # Re-fetch and check for concurrent dispatch (optimistic approach).
+            # No database lock - worst case is duplicate dispatch, but Hatchet
+            # workflows are idempotent so this is acceptable.
+            transcript = await transcripts_controller.get_by_id(config.transcript_id)
+            if transcript and transcript.workflow_run_id:
+                # Another process started a workflow between validation and now
                 try:
                     status = await HatchetClientManager.get_workflow_run_status(
                         transcript.workflow_run_id
                     )
-                    logger.info(
-                        "Old workflow not replayable, starting new",
-                        old_workflow_id=transcript.workflow_run_id,
-                        old_status=status.value,
-                    )
-                except NotFoundException:
-                    # Workflow deleted from Hatchet but ID still in DB
-                    logger.info(
-                        "Old workflow not found in Hatchet, starting new",
-                        old_workflow_id=transcript.workflow_run_id,
-                    )
+                    if status in (V1TaskStatus.RUNNING, V1TaskStatus.QUEUED):
+                        logger.info(
+                            "Concurrent workflow detected, skipping dispatch",
+                            workflow_id=transcript.workflow_run_id,
+                        )
+                        return None
+                except ApiException:
+                    # Workflow might be gone (404) or API issue - proceed with new workflow
+                    pass
 
-        # Force: cancel old workflow if exists
-        if force and transcript and transcript.workflow_run_id:
-            try:
-                await HatchetClientManager.cancel_workflow(transcript.workflow_run_id)
-                logger.info(
-                    "Cancelled old workflow (--force)",
-                    workflow_id=transcript.workflow_run_id,
-                )
-            except NotFoundException:
-                logger.info(
-                    "Old workflow already deleted (--force)",
-                    workflow_id=transcript.workflow_run_id,
-                )
-            await transcripts_controller.update(transcript, {"workflow_run_id": None})
-
-        # Re-fetch and check for concurrent dispatch (optimistic approach).
-        # No database lock - worst case is duplicate dispatch, but Hatchet
-        # workflows are idempotent so this is acceptable.
-        transcript = await transcripts_controller.get_by_id(config.transcript_id)
-        if transcript and transcript.workflow_run_id:
-            # Another process started a workflow between validation and now
-            try:
-                status = await HatchetClientManager.get_workflow_run_status(
-                    transcript.workflow_run_id
-                )
-                if status in (V1TaskStatus.RUNNING, V1TaskStatus.QUEUED):
-                    logger.info(
-                        "Concurrent workflow detected, skipping dispatch",
-                        workflow_id=transcript.workflow_run_id,
-                    )
-                    return None
-            except ApiException:
-                # Workflow might be gone (404) or API issue - proceed with new workflow
-                pass
-
-        workflow_id = await HatchetClientManager.start_workflow(
-            workflow_name="DiarizationPipeline",
-            input_data={
-                "recording_id": config.recording_id,
-                "tracks": [{"s3_key": k} for k in config.track_keys],
-                "bucket_name": config.bucket_name,
-                "transcript_id": config.transcript_id,
-                "room_id": config.room_id,
-            },
-            additional_metadata={
-                "transcript_id": config.transcript_id,
-                "recording_id": config.recording_id,
-                "daily_recording_id": config.recording_id,
-            },
-        )
-
-        if transcript:
-            await transcripts_controller.update(
-                transcript, {"workflow_run_id": workflow_id}
+            workflow_id = await HatchetClientManager.start_workflow(
+                workflow_name="DiarizationPipeline",
+                input_data={
+                    "recording_id": config.recording_id,
+                    "tracks": [{"s3_key": k} for k in config.track_keys],
+                    "bucket_name": config.bucket_name,
+                    "transcript_id": config.transcript_id,
+                    "room_id": config.room_id,
+                },
+                additional_metadata={
+                    "transcript_id": config.transcript_id,
+                    "recording_id": config.recording_id,
+                    "daily_recording_id": config.recording_id,
+                },
             )
 
-        logger.info("Hatchet workflow dispatched", workflow_id=workflow_id)
-        return None
+            if transcript:
+                await transcripts_controller.update(
+                    transcript, {"workflow_run_id": workflow_id}
+                )
 
+            logger.info("Hatchet workflow dispatched", workflow_id=workflow_id)
+            return None
+
+        # Celery pipeline (durable workflows disabled)
+        return task_pipeline_multitrack_process.delay(
+            transcript_id=config.transcript_id,
+            bucket_name=config.bucket_name,
+            track_keys=config.track_keys,
+        )
     elif isinstance(config, FileProcessingConfig):
         return task_pipeline_file_process.delay(transcript_id=config.transcript_id)
     else:

server/reflector/settings.py

@@ -1,7 +1,7 @@
 from pydantic.types import PositiveInt
 from pydantic_settings import BaseSettings, SettingsConfigDict
 
-from reflector.schemas.platform import DAILY_PLATFORM, Platform
+from reflector.schemas.platform import WHEREBY_PLATFORM, Platform
 from reflector.utils.string import NonEmptyString
 
 
@@ -49,7 +49,6 @@ class Settings(BaseSettings):
     TRANSCRIPT_STORAGE_AWS_REGION: str = "us-east-1"
     TRANSCRIPT_STORAGE_AWS_ACCESS_KEY_ID: str | None = None
     TRANSCRIPT_STORAGE_AWS_SECRET_ACCESS_KEY: str | None = None
-    TRANSCRIPT_STORAGE_AWS_ENDPOINT_URL: str | None = None
 
     # Platform-specific recording storage (follows {PREFIX}_STORAGE_AWS_{CREDENTIAL} pattern)
     # Whereby storage configuration
@@ -85,7 +84,9 @@ class Settings(BaseSettings):
     )
 
     # Diarization
-    # backend: modal — HTTP API client (works with Modal.com OR self-hosted gpu/self_hosted/)
+    # backends:
+    #   - pyannote: in-process model loading (no HTTP, runs in same process)
+    #   - modal: HTTP API client (works with Modal.com OR self-hosted gpu/self_hosted/)
     DIARIZATION_ENABLED: bool = True
     DIARIZATION_BACKEND: str = "modal"
     DIARIZATION_URL: str | None = None
@@ -94,6 +95,9 @@ class Settings(BaseSettings):
     # Diarization: modal backend
     DIARIZATION_MODAL_API_KEY: str | None = None
 
+    # Diarization: local pyannote.audio
+    DIARIZATION_PYANNOTE_AUTH_TOKEN: str | None = None
+
     # Audio Padding (Modal.com backend)
     PADDING_URL: str | None = None
     PADDING_MODAL_API_KEY: str | None = None
@@ -151,7 +155,7 @@ class Settings(BaseSettings):
         None  # Webhook UUID for this environment. Not used by production code
     )
     # Platform Configuration
-    DEFAULT_VIDEO_PLATFORM: Platform = DAILY_PLATFORM
+    DEFAULT_VIDEO_PLATFORM: Platform = WHEREBY_PLATFORM
 
     # Zulip integration
     ZULIP_REALM: str | None = None

server/reflector/storage/storage_aws.py

@@ -53,7 +53,6 @@ class AwsStorage(Storage):
         aws_access_key_id: str | None = None,
         aws_secret_access_key: str | None = None,
         aws_role_arn: str | None = None,
-        aws_endpoint_url: str | None = None,
     ):
         if not aws_bucket_name:
             raise ValueError("Storage `aws_storage` require `aws_bucket_name`")
@@ -74,26 +73,17 @@ class AwsStorage(Storage):
         self._access_key_id = aws_access_key_id
         self._secret_access_key = aws_secret_access_key
         self._role_arn = aws_role_arn
-        self._endpoint_url = aws_endpoint_url
 
         self.aws_folder = ""
         if "/" in aws_bucket_name:
             self._bucket_name, self.aws_folder = aws_bucket_name.split("/", 1)
-
-        config_kwargs: dict = {"retries": {"max_attempts": 3, "mode": "adaptive"}}
-        if aws_endpoint_url:
-            config_kwargs["s3"] = {"addressing_style": "path"}
-        self.boto_config = Config(**config_kwargs)
-
+        self.boto_config = Config(retries={"max_attempts": 3, "mode": "adaptive"})
         self.session = aioboto3.Session(
             aws_access_key_id=aws_access_key_id,
             aws_secret_access_key=aws_secret_access_key,
             region_name=aws_region,
         )
-        if aws_endpoint_url:
-            self.base_url = f"{aws_endpoint_url}/{self._bucket_name}/"
-        else:
-            self.base_url = f"https://{self._bucket_name}.s3.amazonaws.com/"
+        self.base_url = f"https://{self._bucket_name}.s3.amazonaws.com/"
 
     # Implement credential properties
     @property
@@ -149,9 +139,7 @@ class AwsStorage(Storage):
         s3filename = f"{folder}/{filename}" if folder else filename
         logger.info(f"Uploading {filename} to S3 {actual_bucket}/{folder}")
 
-        async with self.session.client(
-            "s3", config=self.boto_config, endpoint_url=self._endpoint_url
-        ) as client:
+        async with self.session.client("s3", config=self.boto_config) as client:
             if isinstance(data, bytes):
                 await client.put_object(Bucket=actual_bucket, Key=s3filename, Body=data)
             else:
@@ -174,9 +162,7 @@ class AwsStorage(Storage):
         actual_bucket = bucket or self._bucket_name
         folder = self.aws_folder
         s3filename = f"{folder}/{filename}" if folder else filename
-        async with self.session.client(
-            "s3", config=self.boto_config, endpoint_url=self._endpoint_url
-        ) as client:
+        async with self.session.client("s3", config=self.boto_config) as client:
             presigned_url = await client.generate_presigned_url(
                 operation,
                 Params={"Bucket": actual_bucket, "Key": s3filename},
@@ -191,9 +177,7 @@ class AwsStorage(Storage):
         folder = self.aws_folder
         logger.info(f"Deleting {filename} from S3 {actual_bucket}/{folder}")
         s3filename = f"{folder}/{filename}" if folder else filename
-        async with self.session.client(
-            "s3", config=self.boto_config, endpoint_url=self._endpoint_url
-        ) as client:
+        async with self.session.client("s3", config=self.boto_config) as client:
             await client.delete_object(Bucket=actual_bucket, Key=s3filename)
 
     @handle_s3_client_errors("download")
@@ -202,9 +186,7 @@ class AwsStorage(Storage):
         folder = self.aws_folder
         logger.info(f"Downloading {filename} from S3 {actual_bucket}/{folder}")
         s3filename = f"{folder}/{filename}" if folder else filename
-        async with self.session.client(
-            "s3", config=self.boto_config, endpoint_url=self._endpoint_url
-        ) as client:
+        async with self.session.client("s3", config=self.boto_config) as client:
             response = await client.get_object(Bucket=actual_bucket, Key=s3filename)
             return await response["Body"].read()
 
@@ -219,9 +201,7 @@ class AwsStorage(Storage):
         logger.info(f"Listing objects from S3 {actual_bucket} with prefix '{s3prefix}'")
 
         keys = []
-        async with self.session.client(
-            "s3", config=self.boto_config, endpoint_url=self._endpoint_url
-        ) as client:
+        async with self.session.client("s3", config=self.boto_config) as client:
             paginator = client.get_paginator("list_objects_v2")
             async for page in paginator.paginate(Bucket=actual_bucket, Prefix=s3prefix):
                 if "Contents" in page:
@@ -247,9 +227,7 @@ class AwsStorage(Storage):
         folder = self.aws_folder
         logger.info(f"Streaming {filename} from S3 {actual_bucket}/{folder}")
         s3filename = f"{folder}/{filename}" if folder else filename
-        async with self.session.client(
-            "s3", config=self.boto_config, endpoint_url=self._endpoint_url
-        ) as client:
+        async with self.session.client("s3", config=self.boto_config) as client:
             await client.download_fileobj(
                 Bucket=actual_bucket, Key=s3filename, Fileobj=fileobj
             )

server/reflector/video_platforms/daily.py

@@ -129,6 +129,10 @@ class DailyClient(VideoPlatformClient):
         """Get room presence/session data for a Daily.co room."""
         return await self._api_client.get_room_presence(room_name)
 
+    async def delete_room(self, room_name: str) -> None:
+        """Delete a Daily.co room (idempotent - succeeds even if room doesn't exist)."""
+        return await self._api_client.delete_room(room_name)
+
     async def get_meeting_participants(
         self, meeting_id: str
     ) -> MeetingParticipantsResponse:

server/reflector/views/daily.py

@@ -1,4 +1,6 @@
 import json
+import os
+from datetime import datetime, timezone
 from typing import assert_never
 
 from fastapi import APIRouter, HTTPException, Request
@@ -12,7 +14,10 @@ from reflector.dailyco_api import (
     RecordingReadyEvent,
     RecordingStartedEvent,
 )
+from reflector.dailyco_api.recording_orphans import create_and_log_orphan
+from reflector.db.daily_recording_requests import daily_recording_requests_controller
 from reflector.db.meetings import meetings_controller
+from reflector.db.recordings import Recording, recordings_controller
 from reflector.logger import logger as _logger
 from reflector.settings import settings
 from reflector.video_platforms.factory import create_platform_client
@@ -80,14 +85,7 @@ async def webhook(request: Request):
     try:
         event = event_adapter.validate_python(body_json)
     except Exception as e:
-        err_detail = str(e)
-        if hasattr(e, "errors"):
-            err_detail = f"{err_detail}; errors={e.errors()!r}"
-        logger.error(
-            "Failed to parse webhook event",
-            error=err_detail,
-            body=body.decode(),
-        )
+        logger.error("Failed to parse webhook event", error=str(e), body=body.decode())
         raise HTTPException(status_code=422, detail="Invalid event format")
 
     match event:
@@ -219,10 +217,73 @@ async def _handle_recording_ready(event: RecordingReadyEvent):
 
         track_keys = [t.s3Key for t in tracks if t.type == "audio"]
 
+        # Lookup request
+        match = await daily_recording_requests_controller.find_by_recording_id(
+            recording_id
+        )
+
+        if not match:
+            await create_and_log_orphan(
+                recording_id=recording_id,
+                bucket_name=bucket_name,
+                room_name=room_name,
+                start_ts=event.payload.start_ts,
+                track_keys=track_keys,
+                source="webhook",
+            )
+            return
+
+        meeting_id, _ = match
+
+        # Verify meeting exists
+        meeting = await meetings_controller.get_by_id(meeting_id)
+        if not meeting:
+            logger.error(
+                "Meeting not found (webhook)",
+                recording_id=recording_id,
+                meeting_id=meeting_id,
+            )
+            await create_and_log_orphan(
+                recording_id=recording_id,
+                bucket_name=bucket_name,
+                room_name=room_name,
+                start_ts=event.payload.start_ts,
+                track_keys=track_keys,
+                source="webhook",
+            )
+            return
+
+        # Create recording atomically
+        created = await recordings_controller.try_create_with_meeting(
+            Recording(
+                id=recording_id,
+                bucket_name=bucket_name,
+                object_key=(
+                    os.path.dirname(track_keys[0]) if track_keys else room_name
+                ),
+                recorded_at=datetime.fromtimestamp(
+                    event.payload.start_ts, tz=timezone.utc
+                ),
+                track_keys=track_keys,
+                meeting_id=meeting_id,
+                status="pending",
+            )
+        )
+
+        if not created:
+            # Already created (polling got it first)
+            logger.debug(
+                "Recording already exists (webhook late)",
+                recording_id=recording_id,
+                meeting_id=meeting_id,
+            )
+            return
+
         logger.info(
-            "Raw-tracks recording queuing processing",
+            "Raw-tracks recording queuing processing (webhook)",
             recording_id=recording_id,
             room_name=room_name,
+            meeting_id=meeting_id,
             num_tracks=len(track_keys),
         )
 

server/reflector/views/meetings.py

@@ -1,4 +1,5 @@
 import json
+import logging
 from datetime import datetime, timezone
 from typing import Annotated, Any, Optional
 from uuid import UUID
@@ -9,16 +10,21 @@ from pydantic import BaseModel
 import reflector.auth as auth
 from reflector.dailyco_api import RecordingType
 from reflector.dailyco_api.client import DailyApiError
+from reflector.db.daily_recording_requests import (
+    DailyRecordingRequest,
+    daily_recording_requests_controller,
+)
 from reflector.db.meetings import (
     MeetingConsent,
     meeting_consent_controller,
     meetings_controller,
 )
 from reflector.db.rooms import rooms_controller
-from reflector.logger import logger
 from reflector.utils.string import NonEmptyString
 from reflector.video_platforms.factory import create_platform_client
 
+logger = logging.getLogger(__name__)
+
 router = APIRouter()
 
 
@@ -102,13 +108,6 @@ async def start_recording(
     if not meeting:
         raise HTTPException(status_code=404, detail="Meeting not found")
 
-    log = logger.bind(
-        meeting_id=meeting_id,
-        room_name=meeting.room_name,
-        recording_type=body.type,
-        instance_id=body.instanceId,
-    )
-
     try:
         client = create_platform_client("daily")
         result = await client.start_recording(
@@ -117,9 +116,30 @@ async def start_recording(
             instance_id=body.instanceId,
         )
 
-        log.info(f"Started {body.type} recording via REST API")
+        recording_id = result["id"]
 
-        return {"status": "ok", "result": result}
+        await daily_recording_requests_controller.create(
+            DailyRecordingRequest(
+                recording_id=recording_id,
+                meeting_id=meeting_id,
+                instance_id=body.instanceId,
+                type=body.type,
+                requested_at=datetime.now(timezone.utc),
+            )
+        )
+
+        logger.info(
+            f"Started {body.type} recording via REST API",
+            extra={
+                "meeting_id": meeting_id,
+                "room_name": meeting.room_name,
+                "recording_type": body.type,
+                "instance_id": body.instanceId,
+                "recording_id": recording_id,
+            },
+        )
+
+        return {"status": "ok", "recording_id": recording_id}
 
     except DailyApiError as e:
         # Parse Daily.co error response to detect "has an active stream"
@@ -130,22 +150,42 @@ async def start_recording(
             # "has an active stream" means recording already started by another participant
             # This is SUCCESS from business logic perspective - return 200
             if "has an active stream" in error_info:
-                log.info(
-                    f"{body.type} recording already active (started by another participant)"
+                logger.info(
+                    f"{body.type} recording already active (started by another participant)",
+                    extra={
+                        "meeting_id": meeting_id,
+                        "room_name": meeting.room_name,
+                        "recording_type": body.type,
+                        "instance_id": body.instanceId,
+                    },
                 )
                 return {"status": "already_active", "instanceId": str(body.instanceId)}
         except (json.JSONDecodeError, KeyError):
             pass  # Fall through to error handling
 
         # All other Daily.co API errors
-        log.error(f"Failed to start {body.type} recording", error=str(e))
+        logger.error(
+            f"Failed to start {body.type} recording",
+            extra={
+                "meeting_id": meeting_id,
+                "recording_type": body.type,
+                "error": str(e),
+            },
+        )
         raise HTTPException(
             status_code=500, detail=f"Failed to start recording: {str(e)}"
         )
 
     except Exception as e:
         # Non-Daily.co errors
-        log.error(f"Failed to start {body.type} recording", error=str(e))
+        logger.error(
+            f"Failed to start {body.type} recording",
+            extra={
+                "meeting_id": meeting_id,
+                "recording_type": body.type,
+                "error": str(e),
+            },
+        )
         raise HTTPException(
             status_code=500, detail=f"Failed to start recording: {str(e)}"
         )

server/reflector/views/rooms.py

@@ -20,6 +20,7 @@ from reflector.services.ics_sync import ics_sync_service
 from reflector.settings import settings
 from reflector.utils.url import add_query_param
 from reflector.video_platforms.factory import create_platform_client
+from reflector.worker.process import poll_daily_room_presence_task
 from reflector.worker.webhook import test_webhook
 
 logger = logging.getLogger(__name__)
@@ -365,6 +366,53 @@ async def rooms_create_meeting(
     return meeting
 
 
+@router.post("/rooms/{room_name}/meetings/{meeting_id}/joined")
+async def rooms_joined_meeting(
+    room_name: str,
+    meeting_id: str,
+):
+    """Trigger presence poll (ideally when user actually joins meeting in Daily iframe)"""
+    room = await rooms_controller.get_by_name(room_name)
+    if not room:
+        raise HTTPException(status_code=404, detail="Room not found")
+
+    meeting = await meetings_controller.get_by_id(meeting_id, room=room)
+    if not meeting:
+        raise HTTPException(status_code=404, detail="Meeting not found")
+
+    if meeting.platform == "daily":
+        poll_daily_room_presence_task.delay(meeting_id)
+
+    return {"status": "ok"}
+
+
+@router.post("/rooms/{room_name}/meetings/{meeting_id}/leave")
+async def rooms_leave_meeting(
+    room_name: str,
+    meeting_id: str,
+    delay_seconds: int = 2,
+):
+    """Trigger presence recheck when user leaves meeting (e.g., tab close/navigation).
+
+    Queues presence poll with optional delay to allow Daily.co to detect disconnect.
+    """
+    room = await rooms_controller.get_by_name(room_name)
+    if not room:
+        raise HTTPException(status_code=404, detail="Room not found")
+
+    meeting = await meetings_controller.get_by_id(meeting_id, room=room)
+    if not meeting:
+        raise HTTPException(status_code=404, detail="Meeting not found")
+
+    if meeting.platform == "daily":
+        poll_daily_room_presence_task.apply_async(
+            args=[meeting_id],
+            countdown=delay_seconds,
+        )
+
+    return {"status": "ok"}
+
+
 @router.post("/rooms/{room_id}/webhook/test", response_model=WebhookTestResult)
 async def rooms_test_webhook(
     room_id: str,

server/reflector/views/transcripts_upload.py

@@ -5,7 +5,7 @@ from fastapi import APIRouter, Depends, HTTPException, UploadFile
 from pydantic import BaseModel
 
 import reflector.auth as auth
-from reflector.db.transcripts import SourceKind, transcripts_controller
+from reflector.db.transcripts import transcripts_controller
 from reflector.pipelines.main_file_pipeline import task_pipeline_file_process
 
 router = APIRouter()
@@ -88,10 +88,8 @@ async def transcript_record_upload(
     finally:
         container.close()
 
-    # set the status to "uploaded" and mark as file source
-    await transcripts_controller.update(
-        transcript, {"status": "uploaded", "source_kind": SourceKind.FILE}
-    )
+    # set the status to "uploaded"
+    await transcripts_controller.update(transcript, {"status": "uploaded"})
 
     # launch a background task to process the file
     task_pipeline_file_process.delay(transcript_id=transcript_id)

server/reflector/views/user_websocket.py

@@ -1,6 +1,6 @@
 from typing import Optional
 
-from fastapi import APIRouter, WebSocket, WebSocketDisconnect
+from fastapi import APIRouter, WebSocket
 
 from reflector.auth.auth_jwt import JWTAuth  # type: ignore
 from reflector.db.users import user_controller
@@ -60,8 +60,6 @@ async def user_events_websocket(websocket: WebSocket):
     try:
         while True:
             await websocket.receive()
-    except (RuntimeError, WebSocketDisconnect):
-        pass
     finally:
         if room_id:
             await ws_manager.remove_user_from_room(room_id, websocket)

server/reflector/worker/process.py

@@ -1,6 +1,5 @@
 import json
 import os
-import re
 from datetime import datetime, timezone
 from typing import List, Literal
 from urllib.parse import unquote
@@ -13,10 +12,12 @@ from celery.utils.log import get_task_logger
 from pydantic import ValidationError
 
 from reflector.dailyco_api import FinishedRecordingResponse, RecordingResponse
+from reflector.dailyco_api.recording_orphans import create_and_log_orphan
 from reflector.db.daily_participant_sessions import (
     DailyParticipantSession,
     daily_participant_sessions_controller,
 )
+from reflector.db.daily_recording_requests import daily_recording_requests_controller
 from reflector.db.meetings import meetings_controller
 from reflector.db.recordings import Recording, recordings_controller
 from reflector.db.rooms import rooms_controller
@@ -27,6 +28,9 @@ from reflector.db.transcripts import (
 from reflector.hatchet.client import HatchetClientManager
 from reflector.pipelines.main_file_pipeline import task_pipeline_file_process
 from reflector.pipelines.main_live_pipeline import asynctask
+from reflector.pipelines.main_multitrack_pipeline import (
+    task_pipeline_multitrack_process,
+)
 from reflector.pipelines.topic_processing import EmptyPipeline
 from reflector.processors import AudioFileWriterProcessor
 from reflector.processors.audio_waveform_processor import AudioWaveformProcessor
@@ -227,79 +231,44 @@ async def _process_multitrack_recording_inner(
     recording_start_ts: int,
 ):
     """
-    Process multitrack recording (first time or reprocessing).
+    Process multitrack recording.
 
-    For first processing (webhook/polling):
-    - Uses recording_start_ts for time-based meeting matching (no instanceId available)
-
-    For reprocessing:
-    - Uses recording.meeting_id directly (already linked during first processing)
-    - recording_start_ts is ignored
+    Recording must already exist with meeting_id set (created by webhook/polling before queueing).
     """
 
-    tz = timezone.utc
-    recorded_at = datetime.now(tz)
-    try:
-        if track_keys:
-            folder = os.path.basename(os.path.dirname(track_keys[0]))
-            ts_match = re.search(r"(\d{14})$", folder)
-            if ts_match:
-                ts = ts_match.group(1)
-                recorded_at = datetime.strptime(ts, "%Y%m%d%H%M%S").replace(tzinfo=tz)
-    except Exception as e:
-        logger.warning(
-            f"Could not parse recorded_at from keys, using now() {recorded_at}",
-            e,
-            exc_info=True,
-        )
-
-    # Check if recording already exists (reprocessing path)
+    # Get recording (must exist - created by webhook/polling)
     recording = await recordings_controller.get_by_id(recording_id)
 
-    if recording and recording.meeting_id:
-        # Reprocessing: recording exists with meeting already linked
-        meeting = await meetings_controller.get_by_id(recording.meeting_id)
-        if not meeting:
-            logger.error(
-                "Reprocessing: meeting not found for recording - skipping",
-                meeting_id=recording.meeting_id,
-                recording_id=recording_id,
-            )
-            return
+    if not recording:
+        logger.error(
+            "Recording not found - should have been created by webhook/polling",
+            recording_id=recording_id,
+        )
+        return
 
-        logger.info(
-            "Reprocessing: using existing recording.meeting_id",
+    if not recording.meeting_id:
+        logger.error(
+            "Recording has no meeting_id - orphan should not be queued",
             recording_id=recording_id,
-            meeting_id=meeting.id,
-            room_name=daily_room_name,
         )
-    else:
-        # First processing: recording doesn't exist, need time-based matching
-        # (Daily.co doesn't return instanceId in API, must match by timestamp)
-        recording_start = datetime.fromtimestamp(recording_start_ts, tz=timezone.utc)
-        meeting = await meetings_controller.get_by_room_name_and_time(
-            room_name=daily_room_name,
-            recording_start=recording_start,
-            time_window_hours=168,  # 1 week
-        )
-        if not meeting:
-            logger.error(
-                "Raw-tracks: no meeting found within 1-week window (time-based match) - skipping",
-                recording_id=recording_id,
-                room_name=daily_room_name,
-                recording_start_ts=recording_start_ts,
-                recording_start=recording_start.isoformat(),
-            )
-            return  # Skip processing, will retry on next poll
-        logger.info(
-            "First processing: found meeting via time-based matching",
-            meeting_id=meeting.id,
-            room_name=daily_room_name,
+        return
+
+    # Get meeting
+    meeting = await meetings_controller.get_by_id(recording.meeting_id)
+    if not meeting:
+        logger.error(
+            "Meeting not found for recording",
+            meeting_id=recording.meeting_id,
             recording_id=recording_id,
-            time_delta_seconds=abs(
-                (meeting.start_date - recording_start).total_seconds()
-            ),
         )
+        return
+
+    logger.info(
+        "Processing multitrack recording",
+        recording_id=recording_id,
+        meeting_id=meeting.id,
+        room_name=daily_room_name,
+    )
 
     room_name_base = extract_base_room_name(daily_room_name)
 
@@ -307,33 +276,6 @@ async def _process_multitrack_recording_inner(
     if not room:
         raise Exception(f"Room not found: {room_name_base}")
 
-    if not recording:
-        # Create recording (only happens during first processing)
-        object_key_dir = os.path.dirname(track_keys[0]) if track_keys else ""
-        recording = await recordings_controller.create(
-            Recording(
-                id=recording_id,
-                bucket_name=bucket_name,
-                object_key=object_key_dir,
-                recorded_at=recorded_at,
-                meeting_id=meeting.id,
-                track_keys=track_keys,
-            )
-        )
-    elif not recording.meeting_id:
-        # Recording exists but meeting_id is null (failed first processing)
-        # Update with meeting from time-based matching
-        await recordings_controller.set_meeting_id(
-            recording_id=recording.id,
-            meeting_id=meeting.id,
-        )
-        recording.meeting_id = meeting.id
-        logger.info(
-            "Updated existing recording with meeting_id",
-            recording_id=recording.id,
-            meeting_id=meeting.id,
-        )
-
     transcript = await transcripts_controller.get_by_recording_id(recording.id)
     if not transcript:
         transcript = await transcripts_controller.add(
@@ -348,29 +290,49 @@ async def _process_multitrack_recording_inner(
             room_id=room.id,
         )
 
-    # Multitrack processing always uses Hatchet (no Celery fallback)
-    workflow_id = await HatchetClientManager.start_workflow(
-        workflow_name="DiarizationPipeline",
-        input_data={
-            "recording_id": recording_id,
-            "tracks": [{"s3_key": k} for k in filter_cam_audio_tracks(track_keys)],
-            "bucket_name": bucket_name,
-            "transcript_id": transcript.id,
-            "room_id": room.id,
-        },
-        additional_metadata={
-            "transcript_id": transcript.id,
-            "recording_id": recording_id,
-            "daily_recording_id": recording_id,
-        },
-    )
-    logger.info(
-        "Started Hatchet workflow",
-        workflow_id=workflow_id,
-        transcript_id=transcript.id,
-    )
+    use_celery = room and room.use_celery
+    use_hatchet = not use_celery
 
-    await transcripts_controller.update(transcript, {"workflow_run_id": workflow_id})
+    if use_celery:
+        logger.info(
+            "Room uses legacy Celery processing",
+            room_id=room.id,
+            transcript_id=transcript.id,
+        )
+
+    if use_hatchet:
+        workflow_id = await HatchetClientManager.start_workflow(
+            workflow_name="DiarizationPipeline",
+            input_data={
+                "recording_id": recording_id,
+                "tracks": [{"s3_key": k} for k in filter_cam_audio_tracks(track_keys)],
+                "bucket_name": bucket_name,
+                "transcript_id": transcript.id,
+                "room_id": room.id,
+            },
+            additional_metadata={
+                "transcript_id": transcript.id,
+                "recording_id": recording_id,
+                "daily_recording_id": recording_id,
+            },
+        )
+        logger.info(
+            "Started Hatchet workflow",
+            workflow_id=workflow_id,
+            transcript_id=transcript.id,
+        )
+
+        await transcripts_controller.update(
+            transcript, {"workflow_run_id": workflow_id}
+        )
+        return
+
+    # Celery pipeline (runs when durable workflows disabled)
+    task_pipeline_multitrack_process.delay(
+        transcript_id=transcript.id,
+        bucket_name=bucket_name,
+        track_keys=filter_cam_audio_tracks(track_keys),
+    )
 
 
 @shared_task
@@ -499,7 +461,7 @@ async def store_cloud_recording(
     Store cloud recording reference in meeting table.
 
     Common function for both webhook and polling code paths.
-    Uses time-based matching to handle duplicate room_name values.
+    Uses direct recording_id lookup via daily_recording_requests table.
 
     Args:
         recording_id: Daily.co recording ID
@@ -512,155 +474,170 @@ async def store_cloud_recording(
     Returns:
         True if stored, False if skipped/failed
     """
-    recording_start = datetime.fromtimestamp(start_ts, tz=timezone.utc)
+    # Lookup request
+    match = await daily_recording_requests_controller.find_by_recording_id(recording_id)
 
-    meeting = await meetings_controller.get_by_room_name_and_time(
-        room_name=room_name,
-        recording_start=recording_start,
-        time_window_hours=168,  # 1 week
-    )
-
-    if not meeting:
-        logger.warning(
-            f"Cloud recording ({source}): no meeting found within 1-week window",
+    if not match:
+        # ORPHAN: No request found (pre-migration recording or failed request creation)
+        await create_and_log_orphan(
             recording_id=recording_id,
+            bucket_name="",
             room_name=room_name,
-            recording_start_ts=start_ts,
-            recording_start=recording_start.isoformat(),
+            start_ts=start_ts,
+            track_keys=None,
+            source=source,
         )
         return False
 
+    meeting_id, _ = match
+
     success = await meetings_controller.set_cloud_recording_if_missing(
-        meeting_id=meeting.id,
+        meeting_id=meeting_id,
         s3_key=s3_key,
         duration=duration,
     )
 
     if not success:
         logger.debug(
-            f"Cloud recording ({source}): already set (race lost)",
+            f"Cloud recording ({source}): already set (stop/restart?)",
             recording_id=recording_id,
             room_name=room_name,
-            meeting_id=meeting.id,
+            meeting_id=meeting_id,
         )
         return False
 
     logger.info(
-        f"Cloud recording stored via {source} (time-based match)",
-        meeting_id=meeting.id,
+        f"Cloud recording stored via {source}",
+        meeting_id=meeting_id,
         recording_id=recording_id,
         s3_key=s3_key,
         duration=duration,
-        time_delta_seconds=abs((meeting.start_date - recording_start).total_seconds()),
     )
     return True
 
 
 async def _poll_cloud_recordings(cloud_recordings: List[FinishedRecordingResponse]):
-    """
-    Store cloud recordings missing from meeting table via polling.
+    """Process cloud recordings (database deduplication, worker-agnostic).
 
-    Uses time-based matching via store_cloud_recording().
+    Cloud recordings stored in meeting.daily_composed_video_s3_key, not recording table.
+    Only first cloud recording per meeting is kept (existing behavior).
     """
     if not cloud_recordings:
         return
 
-    stored_count = 0
-    for recording in cloud_recordings:
-        # Extract S3 key from recording (cloud recordings use s3key field)
-        s3_key = recording.s3key or (recording.s3.key if recording.s3 else None)
-        if not s3_key:
-            logger.warning(
-                "Cloud recording: missing S3 key",
-                recording_id=recording.id,
-                room_name=recording.room_name,
+    for rec in cloud_recordings:
+        # Lookup request
+        match = await daily_recording_requests_controller.find_by_recording_id(rec.id)
+
+        if not match:
+            await create_and_log_orphan(
+                recording_id=rec.id,
+                bucket_name="",
+                room_name=rec.room_name,
+                start_ts=rec.start_ts,
+                track_keys=None,
+                source="polling",
             )
             continue
 
-        stored = await store_cloud_recording(
-            recording_id=recording.id,
-            room_name=recording.room_name,
-            s3_key=s3_key,
-            duration=recording.duration,
-            start_ts=recording.start_ts,
-            source="polling",
-        )
-        if stored:
-            stored_count += 1
+        meeting_id, _ = match
 
-    logger.info(
-        "Cloud recording polling complete",
-        total=len(cloud_recordings),
-        stored=stored_count,
-    )
+        if not rec.s3key:
+            logger.error("Cloud recording missing s3_key", recording_id=rec.id)
+            continue
+
+        # Store in meeting table (atomic, only if not already set)
+        success = await meetings_controller.set_cloud_recording_if_missing(
+            meeting_id=meeting_id,
+            s3_key=rec.s3key,
+            duration=rec.duration,
+        )
+
+        if success:
+            logger.info(
+                "Stored cloud recording", recording_id=rec.id, meeting_id=meeting_id
+            )
+        else:
+            logger.warning(
+                "Cloud recording already exists for meeting (stop/restart?)",
+                recording_id=rec.id,
+                meeting_id=meeting_id,
+            )
 
 
 async def _poll_raw_tracks_recordings(
     raw_tracks_recordings: List[FinishedRecordingResponse],
-    bucket_name: str,
-):
-    """Queue raw-tracks recordings missing from DB (existing logic)."""
+    bucket_name: NonEmptyString,
+) -> None:
+    """Process raw-tracks (database deduplication, worker-agnostic)."""
     if not raw_tracks_recordings:
         return
 
-    recording_ids = [rec.id for rec in raw_tracks_recordings]
-    existing_recordings = await recordings_controller.get_by_ids(recording_ids)
-    existing_ids = {rec.id for rec in existing_recordings}
+    for rec in raw_tracks_recordings:
+        # Lookup request FIRST (before any DB writes)
+        match = await daily_recording_requests_controller.find_by_recording_id(rec.id)
 
-    missing_recordings = [
-        rec for rec in raw_tracks_recordings if rec.id not in existing_ids
-    ]
-
-    if not missing_recordings:
-        logger.debug(
-            "All raw-tracks recordings already in DB",
-            api_count=len(raw_tracks_recordings),
-            existing_count=len(existing_recordings),
-        )
-        return
-
-    logger.info(
-        "Found raw-tracks recordings missing from DB",
-        missing_count=len(missing_recordings),
-        total_api_count=len(raw_tracks_recordings),
-        existing_count=len(existing_recordings),
-    )
-
-    for recording in missing_recordings:
-        if not recording.tracks:
-            logger.warning(
-                "Finished raw-tracks recording has no tracks (no audio captured)",
-                recording_id=recording.id,
-                room_name=recording.room_name,
+        if not match:
+            await create_and_log_orphan(
+                recording_id=rec.id,
+                bucket_name=bucket_name,
+                room_name=rec.room_name,
+                start_ts=rec.start_ts,
+                track_keys=[t.s3Key for t in rec.tracks if t.type == "audio"],
+                source="polling",
             )
             continue
 
-        track_keys = [t.s3Key for t in recording.tracks if t.type == "audio"]
+        meeting_id, _ = match
 
-        if not track_keys:
-            logger.warning(
-                "No audio tracks found in raw-tracks recording",
-                recording_id=recording.id,
-                room_name=recording.room_name,
-                total_tracks=len(recording.tracks),
+        # Verify meeting exists
+        meeting = await meetings_controller.get_by_id(meeting_id)
+        if not meeting:
+            logger.error(
+                "Meeting not found", recording_id=rec.id, meeting_id=meeting_id
+            )
+            await create_and_log_orphan(
+                recording_id=rec.id,
+                bucket_name=bucket_name,
+                room_name=rec.room_name,
+                start_ts=rec.start_ts,
+                track_keys=[t.s3Key for t in rec.tracks if t.type == "audio"],
+                source="polling",
             )
             continue
 
-        logger.info(
-            "Queueing missing raw-tracks recording for processing",
-            recording_id=recording.id,
-            room_name=recording.room_name,
-            track_count=len(track_keys),
+        # DEDUPLICATION: Atomically create recording (single operation, no race window)
+        # ON CONFLICT → concurrent poller already got it, skip entire logic
+        track_keys = [t.s3Key for t in rec.tracks if t.type == "audio"]
+
+        created = await recordings_controller.try_create_with_meeting(
+            Recording(
+                id=rec.id,
+                bucket_name=bucket_name,
+                object_key=os.path.dirname(track_keys[0]) if track_keys else "",
+                recorded_at=datetime.fromtimestamp(rec.start_ts, tz=timezone.utc),
+                track_keys=track_keys,
+                meeting_id=meeting_id,  # Set at creation (constraint-safe)
+                status="pending",
+            )
         )
 
+        if not created:
+            # Conflict: another poller already created/queued this
+            # Skip all remaining logic (match already done by winner)
+            continue
+
+        # Only winner reaches here - queue processing (works with Celery or Hatchet)
         process_multitrack_recording.delay(
+            recording_id=rec.id,
+            daily_room_name=rec.room_name,
+            recording_start_ts=rec.start_ts,
             bucket_name=bucket_name,
-            daily_room_name=recording.room_name,
-            recording_id=recording.id,
             track_keys=track_keys,
-            recording_start_ts=recording.start_ts,
         )
 
+        logger.info("Queued recording", recording_id=rec.id, meeting_id=meeting_id)
+
 
 async def poll_daily_room_presence(meeting_id: str) -> None:
     """Poll Daily.co room presence and reconcile with DB sessions. New presence is added, old presence is marked as closed.
@@ -822,15 +799,47 @@ async def process_meetings():
                     end_date = end_date.replace(tzinfo=timezone.utc)
 
                 client = create_platform_client(meeting.platform)
-                room_sessions = await client.get_room_sessions(meeting.room_name)
+                has_active_sessions = False
+                has_had_sessions = False
 
-                has_active_sessions = bool(
-                    room_sessions and any(s.ended_at is None for s in room_sessions)
-                )
-                has_had_sessions = bool(room_sessions)
-                logger_.info(
-                    f"has_active_sessions={has_active_sessions}, has_had_sessions={has_had_sessions}"
-                )
+                if meeting.platform == "daily":
+                    try:
+                        presence = await client.get_room_presence(meeting.room_name)
+                        has_active_sessions = presence.total_count > 0
+
+                        room_sessions = await client.get_room_sessions(
+                            meeting.room_name
+                        )
+                        has_had_sessions = bool(room_sessions)
+
+                        logger_.info(
+                            "Daily.co presence check",
+                            has_active_sessions=has_active_sessions,
+                            has_had_sessions=has_had_sessions,
+                            presence_count=presence.total_count,
+                        )
+                    except Exception:
+                        logger_.warning(
+                            "Daily.co presence API failed, falling back to DB sessions",
+                            exc_info=True,
+                        )
+                        room_sessions = await client.get_room_sessions(
+                            meeting.room_name
+                        )
+                        has_active_sessions = bool(
+                            room_sessions
+                            and any(s.ended_at is None for s in room_sessions)
+                        )
+                        has_had_sessions = bool(room_sessions)
+                else:
+                    room_sessions = await client.get_room_sessions(meeting.room_name)
+                    has_active_sessions = bool(
+                        room_sessions and any(s.ended_at is None for s in room_sessions)
+                    )
+                    has_had_sessions = bool(room_sessions)
+                    logger_.info(
+                        f"has_active_sessions={has_active_sessions}, has_had_sessions={has_had_sessions}"
+                    )
 
                 if has_active_sessions:
                     logger_.debug("Meeting still has active sessions, keep it")
@@ -849,7 +858,20 @@ async def process_meetings():
                     await meetings_controller.update_meeting(
                         meeting.id, is_active=False
                     )
-                    logger_.info("Meeting is deactivated")
+                    logger_.info("Meeting deactivated in database")
+
+                    if meeting.platform == "daily":
+                        try:
+                            await client.delete_room(meeting.room_name)
+                            logger_.info(
+                                "Daily.co room deleted", room_name=meeting.room_name
+                            )
+                        except Exception:
+                            logger_.warning(
+                                "Failed to delete Daily.co room",
+                                room_name=meeting.room_name,
+                                exc_info=True,
+                            )
 
                 processed_count += 1
 
@@ -1049,43 +1071,66 @@ async def reprocess_failed_daily_recordings():
                 )
                 continue
 
-            # Multitrack reprocessing always uses Hatchet (no Celery fallback)
-            if not transcript:
-                logger.warning(
-                    "No transcript for Hatchet reprocessing, skipping",
-                    recording_id=recording.id,
+            use_celery = room and room.use_celery
+            use_hatchet = not use_celery
+
+            if use_hatchet:
+                if not transcript:
+                    logger.warning(
+                        "No transcript for Hatchet reprocessing, skipping",
+                        recording_id=recording.id,
+                    )
+                    continue
+
+                workflow_id = await HatchetClientManager.start_workflow(
+                    workflow_name="DiarizationPipeline",
+                    input_data={
+                        "recording_id": recording.id,
+                        "tracks": [
+                            {"s3_key": k}
+                            for k in filter_cam_audio_tracks(recording.track_keys)
+                        ],
+                        "bucket_name": bucket_name,
+                        "transcript_id": transcript.id,
+                        "room_id": room.id if room else None,
+                    },
+                    additional_metadata={
+                        "transcript_id": transcript.id,
+                        "recording_id": recording.id,
+                        "reprocess": True,
+                    },
+                )
+                await transcripts_controller.update(
+                    transcript, {"workflow_run_id": workflow_id}
                 )
-                continue
 
-            workflow_id = await HatchetClientManager.start_workflow(
-                workflow_name="DiarizationPipeline",
-                input_data={
-                    "recording_id": recording.id,
-                    "tracks": [
-                        {"s3_key": k}
-                        for k in filter_cam_audio_tracks(recording.track_keys)
-                    ],
-                    "bucket_name": bucket_name,
-                    "transcript_id": transcript.id,
-                    "room_id": room.id if room else None,
-                },
-                additional_metadata={
-                    "transcript_id": transcript.id,
-                    "recording_id": recording.id,
-                    "reprocess": True,
-                },
-            )
-            await transcripts_controller.update(
-                transcript, {"workflow_run_id": workflow_id}
-            )
+                logger.info(
+                    "Queued Daily recording for Hatchet reprocessing",
+                    recording_id=recording.id,
+                    workflow_id=workflow_id,
+                    room_name=meeting.room_name,
+                    track_count=len(recording.track_keys),
+                )
+            else:
+                logger.info(
+                    "Queueing Daily recording for Celery reprocessing",
+                    recording_id=recording.id,
+                    room_name=meeting.room_name,
+                    track_count=len(recording.track_keys),
+                    transcript_status=transcript.status if transcript else None,
+                )
 
-            logger.info(
-                "Queued Daily recording for Hatchet reprocessing",
-                recording_id=recording.id,
-                workflow_id=workflow_id,
-                room_name=meeting.room_name,
-                track_count=len(recording.track_keys),
-            )
+                # For reprocessing, pass actual recording time (though it's ignored - see _process_multitrack_recording_inner)
+                # Reprocessing uses recording.meeting_id directly instead of time-based matching
+                recording_start_ts = int(recording.recorded_at.timestamp())
+
+                process_multitrack_recording.delay(
+                    bucket_name=bucket_name,
+                    daily_room_name=meeting.room_name,
+                    recording_id=recording.id,
+                    track_keys=recording.track_keys,
+                    recording_start_ts=recording_start_ts,
+                )
 
             reprocessed_count += 1
 

server/reflector/ws_manager.py

@@ -48,15 +48,7 @@ class RedisPubSubManager:
         if not self.redis_connection:
             await self.connect()
         message = json.dumps(message)
-        try:
-            await self.redis_connection.publish(room_id, message)
-        except RuntimeError:
-            # Celery workers run each task in a new event loop (asyncio.run),
-            # which closes the previous loop. Cached Redis connection is dead.
-            # Reconnect on the current loop and retry.
-            self.redis_connection = None
-            await self.connect()
-            await self.redis_connection.publish(room_id, message)
+        await self.redis_connection.publish(room_id, message)
 
     async def subscribe(self, room_id: str) -> redis.Redis:
         await self.pubsub.subscribe(room_id)

server/scripts/recreate_daily_webhook.py

@@ -15,7 +15,8 @@ from reflector.settings import settings
 
 async def setup_webhook(webhook_url: str):
     """
-    Create Daily.co webhook. Deletes any existing webhooks first, then creates the new one.
+    Create or update Daily.co webhook for this environment using dailyco_api module.
+    Uses DAILY_WEBHOOK_UUID to identify existing webhook.
     """
     if not settings.DAILY_API_KEY:
         print("Error: DAILY_API_KEY not set")
@@ -34,37 +35,79 @@ async def setup_webhook(webhook_url: str):
     ]
 
     async with DailyApiClient(api_key=settings.DAILY_API_KEY) as client:
-        webhooks = await client.list_webhooks()
-        for wh in webhooks:
-            await client.delete_webhook(wh.uuid)
-            print(f"Deleted webhook {wh.uuid}")
+        webhook_uuid = settings.DAILY_WEBHOOK_UUID
 
-        request = CreateWebhookRequest(
-            url=webhook_url,
-            eventTypes=event_types,
-            hmac=settings.DAILY_WEBHOOK_SECRET,
-        )
-        result = await client.create_webhook(request)
-        webhook_uuid = result.uuid
+        if webhook_uuid:
+            print(f"Updating existing webhook {webhook_uuid}...")
+            try:
+                # Note: Daily.co doesn't support PATCH well, so we delete + recreate
+                await client.delete_webhook(webhook_uuid)
+                print(f"Deleted old webhook {webhook_uuid}")
 
-        print(f"✓ Created webhook {webhook_uuid} (state: {result.state})")
-        print(f"  URL: {result.url}")
+                request = CreateWebhookRequest(
+                    url=webhook_url,
+                    eventTypes=event_types,
+                    hmac=settings.DAILY_WEBHOOK_SECRET,
+                )
+                result = await client.create_webhook(request)
 
-        env_file = Path(__file__).parent.parent / ".env"
-        if env_file.exists():
-            lines = env_file.read_text().splitlines()
-            updated = False
-            for i, line in enumerate(lines):
-                if line.startswith("DAILY_WEBHOOK_UUID="):
-                    lines[i] = f"DAILY_WEBHOOK_UUID={webhook_uuid}"
-                    updated = True
-                    break
-            if not updated:
-                lines.append(f"DAILY_WEBHOOK_UUID={webhook_uuid}")
-            env_file.write_text("\n".join(lines) + "\n")
-            print("✓ Saved DAILY_WEBHOOK_UUID to .env")
+                print(
+                    f"✓ Created replacement webhook {result.uuid} (state: {result.state})"
+                )
+                print(f"  URL: {result.url}")
 
-        return 0
+                webhook_uuid = result.uuid
+
+            except Exception as e:
+                if hasattr(e, "response") and e.response.status_code == 404:
+                    print(f"Webhook {webhook_uuid} not found, creating new one...")
+                    webhook_uuid = None  # Fall through to creation
+                else:
+                    print(f"Error updating webhook: {e}")
+                    return 1
+
+        if not webhook_uuid:
+            print("Creating new webhook...")
+            request = CreateWebhookRequest(
+                url=webhook_url,
+                eventTypes=event_types,
+                hmac=settings.DAILY_WEBHOOK_SECRET,
+            )
+            result = await client.create_webhook(request)
+            webhook_uuid = result.uuid
+
+            print(f"✓ Created webhook {webhook_uuid} (state: {result.state})")
+            print(f"  URL: {result.url}")
+            print()
+            print("=" * 60)
+            print("IMPORTANT: Add this to your environment variables:")
+            print("=" * 60)
+            print(f"DAILY_WEBHOOK_UUID: {webhook_uuid}")
+            print("=" * 60)
+            print()
+
+            # Try to write UUID to .env file
+            env_file = Path(__file__).parent.parent / ".env"
+            if env_file.exists():
+                lines = env_file.read_text().splitlines()
+                updated = False
+
+                # Update existing DAILY_WEBHOOK_UUID line or add it
+                for i, line in enumerate(lines):
+                    if line.startswith("DAILY_WEBHOOK_UUID="):
+                        lines[i] = f"DAILY_WEBHOOK_UUID={webhook_uuid}"
+                        updated = True
+                        break
+
+                if not updated:
+                    lines.append(f"DAILY_WEBHOOK_UUID={webhook_uuid}")
+
+                env_file.write_text("\n".join(lines) + "\n")
+                print(f"✓ Also saved to local .env file")
+            else:
+                print(f"⚠ Local .env file not found - please add manually")
+
+            return 0
 
 
 if __name__ == "__main__":
@@ -74,7 +117,11 @@ if __name__ == "__main__":
             "Example: python recreate_daily_webhook.py https://example.com/v1/daily/webhook"
         )
         print()
-        print("Deletes all existing webhooks, then creates a new one.")
+        print("Behavior:")
+        print("  - If DAILY_WEBHOOK_UUID set: Deletes old webhook, creates new one")
+        print(
+            "  - If DAILY_WEBHOOK_UUID empty: Creates new webhook, saves UUID to .env"
+        )
         sys.exit(1)
 
     sys.exit(asyncio.run(setup_webhook(sys.argv[1])))

server/test_daily_api_recordings.py

@@ -0,0 +1,39 @@
+#!/usr/bin/env python3
+"""Test script to fetch Daily.co recordings for a specific room and show raw API response."""
+
+import asyncio
+import json
+
+from reflector.video_platforms.factory import create_platform_client
+
+
+async def main():
+    room_name = "daily-private-igor-20260110042117"
+
+    print(f"\n=== Fetching recordings for room: {room_name} ===\n")
+
+    async with create_platform_client("daily") as client:
+        recordings = await client.list_recordings(room_name=room_name)
+
+    print(f"Found {len(recordings)} recording objects from Daily.co API\n")
+
+    for i, rec in enumerate(recordings, 1):
+        print(f"--- Recording #{i} ---")
+        print(f"ID: {rec.id}")
+        print(f"Room: {rec.room_name}")
+        print(f"Start TS: {rec.start_ts}")
+        print(f"Status: {rec.status}")
+        print(f"Duration: {rec.duration}")
+        print(f"Type: {rec.type}")
+        print(f"Tracks count: {len(rec.tracks)}")
+
+        if rec.tracks:
+            print(f"Tracks:")
+            for j, track in enumerate(rec.tracks, 1):
+                print(f"  Track {j}: {track.s3Key}")
+
+        print(f"\nRaw JSON:\n{json.dumps(rec.model_dump(), indent=2, default=str)}\n")
+
+
+if __name__ == "__main__":
+    asyncio.run(main())

server/tests/conftest.py

@@ -4,7 +4,7 @@ from unittest.mock import patch
 
 import pytest
 
-from reflector.schemas.platform import DAILY_PLATFORM, WHEREBY_PLATFORM
+from reflector.schemas.platform import WHEREBY_PLATFORM
 
 
 @pytest.fixture(scope="session", autouse=True)
@@ -14,7 +14,6 @@ def register_mock_platform():
     from reflector.video_platforms.registry import register_platform
 
     register_platform(WHEREBY_PLATFORM, MockPlatformClient)
-    register_platform(DAILY_PLATFORM, MockPlatformClient)
     yield
 
 

server/tests/test_daily_presence_deactivation.py

@@ -0,0 +1,286 @@
+"""Unit tests for Daily.co presence-based meeting deactivation logic.
+
+Tests the fix for split room race condition by verifying:
+1. Real-time presence checking via Daily.co API
+2. Room deletion when meetings deactivate
+"""
+
+from datetime import datetime, timedelta, timezone
+from unittest.mock import AsyncMock, patch
+
+import pytest
+
+from reflector.dailyco_api.responses import (
+    RoomPresenceParticipant,
+    RoomPresenceResponse,
+)
+from reflector.db.daily_participant_sessions import (
+    DailyParticipantSession,
+    daily_participant_sessions_controller,
+)
+from reflector.db.meetings import meetings_controller
+from reflector.db.rooms import rooms_controller
+from reflector.video_platforms.daily import DailyClient
+
+
+@pytest.fixture
+async def daily_room_and_meeting():
+    """Create test room and meeting for Daily platform."""
+    room = await rooms_controller.add(
+        name="test-daily",
+        user_id="test-user",
+        platform="daily",
+        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,
+    )
+
+    current_time = datetime.now(timezone.utc)
+    end_time = current_time + timedelta(hours=2)
+
+    meeting = await meetings_controller.create(
+        id="test-meeting-id",
+        room_name="test-daily-20260129120000",
+        room_url="https://daily.co/test",
+        host_room_url="https://daily.co/test",
+        start_date=current_time,
+        end_date=end_time,
+        room=room,
+    )
+
+    return room, meeting
+
+
+@pytest.mark.asyncio
+async def test_daily_client_has_delete_room_method():
+    """Verify DailyClient has delete_room method for cleanup."""
+    # Create a mock DailyClient
+    with patch("reflector.dailyco_api.client.DailyApiClient"):
+        from reflector.video_platforms.models import VideoPlatformConfig
+
+        config = VideoPlatformConfig(api_key="test-key", webhook_secret="test-secret")
+        client = DailyClient(config)
+
+        # Verify delete_room method exists
+        assert hasattr(client, "delete_room")
+        assert callable(getattr(client, "delete_room"))
+
+
+@pytest.mark.asyncio
+async def test_get_room_presence_returns_realtime_data(daily_room_and_meeting):
+    """Test that get_room_presence returns real-time participant data."""
+    room, meeting = daily_room_and_meeting
+
+    # Mock Daily.co API response
+    mock_presence = RoomPresenceResponse(
+        total_count=2,
+        data=[
+            RoomPresenceParticipant(
+                room=meeting.room_name,
+                id="session-1",
+                userId="user-1",
+                userName="User One",
+                joinTime="2026-01-29T12:00:00.000Z",
+                duration=120,
+            ),
+            RoomPresenceParticipant(
+                room=meeting.room_name,
+                id="session-2",
+                userId="user-2",
+                userName="User Two",
+                joinTime="2026-01-29T12:05:00.000Z",
+                duration=60,
+            ),
+        ],
+    )
+
+    with patch("reflector.dailyco_api.client.DailyApiClient") as mock_api:
+        from reflector.video_platforms.models import VideoPlatformConfig
+
+        config = VideoPlatformConfig(api_key="test-key", webhook_secret="test-secret")
+        client = DailyClient(config)
+
+        # Mock the API client method
+        client._api_client.get_room_presence = AsyncMock(return_value=mock_presence)
+
+        # Call get_room_presence
+        result = await client.get_room_presence(meeting.room_name)
+
+        # Verify it calls Daily.co API
+        client._api_client.get_room_presence.assert_called_once_with(meeting.room_name)
+
+        # Verify result contains real-time data
+        assert result.total_count == 2
+        assert len(result.data) == 2
+        assert result.data[0].id == "session-1"
+        assert result.data[1].id == "session-2"
+
+
+@pytest.mark.asyncio
+async def test_presence_shows_active_even_when_db_stale(daily_room_and_meeting):
+    """Test that Daily.co presence API is source of truth, not stale DB sessions."""
+    room, meeting = daily_room_and_meeting
+    current_time = datetime.now(timezone.utc)
+
+    # Create stale DB session (left_at=NULL but user actually left)
+    session_id = f"{meeting.id}:stale-user:{int((current_time - timedelta(minutes=5)).timestamp() * 1000)}"
+    await daily_participant_sessions_controller.upsert_joined(
+        DailyParticipantSession(
+            id=session_id,
+            meeting_id=meeting.id,
+            room_id=room.id,
+            session_id="stale-daily-session",
+            user_name="Stale User",
+            user_id="stale-user",
+            joined_at=current_time - timedelta(minutes=5),
+            left_at=None,  # Stale - shows active but user left
+        )
+    )
+
+    # Verify DB shows active session
+    db_sessions = await daily_participant_sessions_controller.get_active_by_meeting(
+        meeting.id
+    )
+    assert len(db_sessions) == 1
+
+    # But Daily.co API shows room is empty
+    mock_presence = RoomPresenceResponse(total_count=0, data=[])
+
+    with patch("reflector.dailyco_api.client.DailyApiClient"):
+        from reflector.video_platforms.models import VideoPlatformConfig
+
+        config = VideoPlatformConfig(api_key="test-key", webhook_secret="test-secret")
+        client = DailyClient(config)
+        client._api_client.get_room_presence = AsyncMock(return_value=mock_presence)
+
+        # Get real-time presence
+        presence = await client.get_room_presence(meeting.room_name)
+
+        # Real-time API shows no participants (truth)
+        assert presence.total_count == 0
+        assert len(presence.data) == 0
+
+        # DB shows 1 participant (stale)
+        assert len(db_sessions) == 1
+
+        # Implementation should trust presence API, not DB
+
+
+@pytest.mark.asyncio
+async def test_meeting_deactivation_logic_with_presence_empty():
+    """Test the core deactivation decision logic when presence shows room empty."""
+    # This tests the logic that will be in process_meetings
+
+    # Simulate: DB shows stale active session
+    has_active_db_sessions = True  # DB is stale
+
+    # Simulate: Daily.co presence API shows room empty
+    presence_count = 0  # Real-time truth
+
+    # Simulate: Meeting has been used before
+    has_had_sessions = True
+
+    # Decision logic (what process_meetings should do):
+    # - If presence API available: trust it
+    # - If presence shows empty AND has_had_sessions: deactivate
+
+    if presence_count == 0 and has_had_sessions:
+        should_deactivate = True
+    else:
+        should_deactivate = False
+
+    assert should_deactivate is True  # Should deactivate despite stale DB
+
+
+@pytest.mark.asyncio
+async def test_meeting_deactivation_logic_with_presence_active():
+    """Test that meetings stay active when presence shows participants."""
+    # Simulate: DB shows no sessions (not yet updated)
+    has_active_db_sessions = False  # DB hasn't caught up
+
+    # Simulate: Daily.co presence API shows active participant
+    presence_count = 1  # Real-time truth
+
+    # Decision logic: presence shows activity, keep meeting active
+    if presence_count > 0:
+        should_deactivate = False
+    else:
+        should_deactivate = True
+
+    assert should_deactivate is False  # Should stay active
+
+
+@pytest.mark.asyncio
+async def test_delete_room_called_on_deactivation(daily_room_and_meeting):
+    """Test that Daily.co room is deleted when meeting deactivates."""
+    room, meeting = daily_room_and_meeting
+
+    with patch("reflector.dailyco_api.client.DailyApiClient"):
+        from reflector.video_platforms.models import VideoPlatformConfig
+
+        config = VideoPlatformConfig(api_key="test-key", webhook_secret="test-secret")
+        client = DailyClient(config)
+
+        # Mock delete_room API call
+        client._api_client.delete_room = AsyncMock()
+
+        # Simulate deactivation - should delete room
+        await client._api_client.delete_room(meeting.room_name)
+
+        # Verify delete was called
+        client._api_client.delete_room.assert_called_once_with(meeting.room_name)
+
+
+@pytest.mark.asyncio
+async def test_delete_room_idempotent_on_404():
+    """Test that room deletion is idempotent (succeeds even if room doesn't exist)."""
+    from reflector.dailyco_api.client import DailyApiClient
+
+    # Create real client to test delete_room logic
+    client = DailyApiClient(api_key="test-key")
+
+    # Mock the HTTP client
+    mock_http_client = AsyncMock()
+    mock_response = AsyncMock()
+    mock_response.status_code = 404  # Room not found
+    mock_http_client.delete = AsyncMock(return_value=mock_response)
+
+    # Mock _get_client to return our mock
+    async def mock_get_client():
+        return mock_http_client
+
+    client._get_client = mock_get_client
+
+    # delete_room should succeed even on 404 (idempotent)
+    await client.delete_room("nonexistent-room")
+
+    # Verify delete was attempted
+    mock_http_client.delete.assert_called_once()
+
+
+@pytest.mark.asyncio
+async def test_api_failure_fallback_to_db_sessions():
+    """Test that system falls back to DB sessions if Daily.co API fails."""
+    # Simulate: Daily.co API throws exception
+    api_exception = Exception("API unavailable")
+
+    # Simulate: DB shows active session
+    has_active_db_sessions = True
+
+    # Decision logic with fallback:
+    try:
+        presence_count = None
+        raise api_exception  # Simulating API failure
+    except Exception:
+        # Fallback: use DB sessions (conservative - don't deactivate if unsure)
+        if has_active_db_sessions:
+            should_deactivate = False
+        else:
+            should_deactivate = True
+
+    assert should_deactivate is False  # Conservative: keep active on API failure

server/tests/test_daily_recording_requests.py

@@ -0,0 +1,258 @@
+from datetime import datetime, timezone
+from uuid import UUID
+
+import pytest
+
+from reflector.db.daily_recording_requests import (
+    DailyRecordingRequest,
+    daily_recording_requests_controller,
+)
+from reflector.db.meetings import Meeting, meetings_controller
+from reflector.db.recordings import Recording, recordings_controller
+from reflector.db.rooms import Room, rooms_controller
+
+
+@pytest.mark.asyncio
+async def test_create_request():
+    """Test creating a recording request."""
+    # Create meeting first
+    room = Room(id="test-room", name="Test Room", slug="test-room", user_id="test-user")
+    await rooms_controller.create(room)
+
+    meeting = Meeting(
+        id="meeting-123",
+        room_name="test-room",
+        start_date=datetime.now(timezone.utc),
+        end_date=None,
+        recording_type="cloud",
+    )
+    await meetings_controller.create(meeting)
+
+    request = DailyRecordingRequest(
+        recording_id="rec-1",
+        meeting_id="meeting-123",
+        instance_id=UUID("a1b2c3d4-e5f6-7890-abcd-ef1234567890"),
+        type="cloud",
+        requested_at=datetime.now(timezone.utc),
+    )
+
+    await daily_recording_requests_controller.create(request)
+
+    result = await daily_recording_requests_controller.find_by_recording_id("rec-1")
+    assert result is not None
+    assert result[0] == "meeting-123"
+    assert result[1] == "cloud"
+
+
+@pytest.mark.asyncio
+async def test_multiple_recordings_same_meeting():
+    """Test stop/restart creates multiple request rows."""
+    # Create room and meeting
+    room = Room(
+        id="test-room-2", name="Test Room 2", slug="test-room-2", user_id="test-user"
+    )
+    await rooms_controller.create(room)
+
+    meeting_id = "meeting-456"
+    meeting = Meeting(
+        id=meeting_id,
+        room_name="test-room-2",
+        start_date=datetime.now(timezone.utc),
+        end_date=None,
+        recording_type="cloud",
+    )
+    await meetings_controller.create(meeting)
+
+    instance_id = UUID("b1c2d3e4-f5a6-7890-abcd-ef1234567890")
+
+    # First recording
+    await daily_recording_requests_controller.create(
+        DailyRecordingRequest(
+            recording_id="rec-1",
+            meeting_id=meeting_id,
+            instance_id=instance_id,
+            type="cloud",
+            requested_at=datetime.now(timezone.utc),
+        )
+    )
+
+    # Stop, then restart (new recording_id, same instance_id)
+    await daily_recording_requests_controller.create(
+        DailyRecordingRequest(
+            recording_id="rec-2",  # DIFFERENT
+            meeting_id=meeting_id,
+            instance_id=instance_id,  # SAME
+            type="cloud",
+            requested_at=datetime.now(timezone.utc),
+        )
+    )
+
+    # Both exist
+    requests = await daily_recording_requests_controller.get_by_meeting_id(meeting_id)
+    assert len(requests) == 2
+    assert {r.recording_id for r in requests} == {"rec-1", "rec-2"}
+
+
+@pytest.mark.asyncio
+async def test_deduplication_via_database():
+    """Test concurrent pollers use database for deduplication."""
+    # Create room and meeting
+    room = Room(
+        id="test-room-3", name="Test Room 3", slug="test-room-3", user_id="test-user"
+    )
+    await rooms_controller.create(room)
+
+    meeting = Meeting(
+        id="meeting-789",
+        room_name="test-room-3",
+        start_date=datetime.now(timezone.utc),
+        end_date=None,
+        recording_type="raw-tracks",
+    )
+    await meetings_controller.create(meeting)
+
+    recording_id = "rec-123"
+
+    # Poller 1
+    created1 = await recordings_controller.try_create_with_meeting(
+        Recording(
+            id=recording_id,
+            bucket_name="test-bucket",
+            object_key="test-key",
+            recorded_at=datetime.now(timezone.utc),
+            meeting_id="meeting-789",
+            status="pending",
+            track_keys=["track1.webm", "track2.webm"],
+        )
+    )
+    assert created1 is True  # First wins
+
+    # Poller 2 (concurrent)
+    created2 = await recordings_controller.try_create_with_meeting(
+        Recording(
+            id=recording_id,
+            bucket_name="test-bucket",
+            object_key="test-key",
+            recorded_at=datetime.now(timezone.utc),
+            meeting_id="meeting-789",
+            status="pending",
+            track_keys=["track1.webm", "track2.webm"],
+        )
+    )
+    assert created2 is False  # Conflict, skip
+
+
+@pytest.mark.asyncio
+async def test_orphan_logged_once():
+    """Test orphan marked once, skipped on re-poll."""
+    # First poll
+    created1 = await recordings_controller.create_orphan(
+        Recording(
+            id="orphan-123",
+            bucket_name="test-bucket",
+            object_key="orphan-key",
+            recorded_at=datetime.now(timezone.utc),
+            meeting_id=None,
+            status="orphan",
+            track_keys=None,
+        )
+    )
+    assert created1 is True
+
+    # Second poll (same orphan discovered again)
+    created2 = await recordings_controller.create_orphan(
+        Recording(
+            id="orphan-123",
+            bucket_name="test-bucket",
+            object_key="orphan-key",
+            recorded_at=datetime.now(timezone.utc),
+            meeting_id=None,
+            status="orphan",
+            track_keys=None,
+        )
+    )
+    assert created2 is False  # Already exists
+
+    # Verify it exists
+    existing = await recordings_controller.get_by_id("orphan-123")
+    assert existing is not None
+    assert existing.status == "orphan"
+
+
+@pytest.mark.asyncio
+async def test_orphan_constraints():
+    """Test orphan invariants are enforced."""
+    # Can't create orphan with meeting_id
+    with pytest.raises(AssertionError, match="meeting_id must be NULL"):
+        await recordings_controller.create_orphan(
+            Recording(
+                id="bad-orphan-1",
+                bucket_name="test",
+                object_key="test",
+                recorded_at=datetime.now(timezone.utc),
+                meeting_id="meeting-123",  # Should be None
+                status="orphan",
+                track_keys=None,
+            )
+        )
+
+    # Can't create orphan with wrong status
+    with pytest.raises(AssertionError, match="status must be 'orphan'"):
+        await recordings_controller.create_orphan(
+            Recording(
+                id="bad-orphan-2",
+                bucket_name="test",
+                object_key="test",
+                recorded_at=datetime.now(timezone.utc),
+                meeting_id=None,
+                status="pending",  # Should be "orphan"
+                track_keys=None,
+            )
+        )
+
+
+@pytest.mark.asyncio
+async def test_try_create_with_meeting_constraints():
+    """Test try_create_with_meeting enforces constraints."""
+    # Create room and meeting
+    room = Room(
+        id="test-room-4", name="Test Room 4", slug="test-room-4", user_id="test-user"
+    )
+    await rooms_controller.create(room)
+
+    meeting = Meeting(
+        id="meeting-999",
+        room_name="test-room-4",
+        start_date=datetime.now(timezone.utc),
+        end_date=None,
+        recording_type="cloud",
+    )
+    await meetings_controller.create(meeting)
+
+    # Can't create with orphan status
+    with pytest.raises(AssertionError, match="use create_orphan"):
+        await recordings_controller.try_create_with_meeting(
+            Recording(
+                id="bad-rec-1",
+                bucket_name="test",
+                object_key="test",
+                recorded_at=datetime.now(timezone.utc),
+                meeting_id="meeting-999",
+                status="orphan",  # Should not be orphan
+                track_keys=None,
+            )
+        )
+
+    # Can't create without meeting_id
+    with pytest.raises(AssertionError, match="meeting_id required"):
+        await recordings_controller.try_create_with_meeting(
+            Recording(
+                id="bad-rec-2",
+                bucket_name="test",
+                object_key="test",
+                recorded_at=datetime.now(timezone.utc),
+                meeting_id=None,  # Should have meeting_id
+                status="pending",
+                track_keys=None,
+            )
+        )

server/tests/test_hatchet_dispatch.py

@@ -255,7 +255,7 @@ async def test_validation_locked_transcript():
 @pytest.mark.usefixtures("setup_database")
 @pytest.mark.asyncio
 async def test_validation_idle_transcript():
-    """Test that validation rejects idle transcripts without recording (file upload not ready)."""
+    """Test that validation rejects idle transcripts (not ready)."""
     from reflector.services.transcript_process import (
         ValidationNotReady,
         validate_transcript_for_processing,
@@ -274,34 +274,6 @@ async def test_validation_idle_transcript():
     assert "not ready" in result.detail.lower()
 
 
-@pytest.mark.usefixtures("setup_database")
-@pytest.mark.asyncio
-async def test_validation_idle_transcript_with_recording_allowed():
-    """Test that validation allows idle transcripts with recording_id (multitrack ready/retry)."""
-    from reflector.services.transcript_process import (
-        ValidationOk,
-        validate_transcript_for_processing,
-    )
-
-    mock_transcript = Transcript(
-        id="test-transcript-id",
-        name="Test",
-        status="idle",
-        source_kind="room",
-        recording_id="test-recording-id",
-    )
-
-    with patch(
-        "reflector.services.transcript_process.task_is_scheduled_or_active"
-    ) as mock_celery_check:
-        mock_celery_check.return_value = False
-
-        result = await validate_transcript_for_processing(mock_transcript)
-
-    assert isinstance(result, ValidationOk)
-    assert result.recording_id == "test-recording-id"
-
-
 @pytest.mark.usefixtures("setup_database")
 @pytest.mark.asyncio
 async def test_prepare_multitrack_config():

server/tests/test_recording_request_flow.py

@@ -0,0 +1,300 @@
+"""
+Integration tests for recording request flow.
+
+These tests verify the end-to-end flow of:
+1. Starting a recording (creates request)
+2. Webhook/polling discovering recording (matches via request)
+3. Recording processing (uses existing meeting_id)
+"""
+
+from datetime import datetime, timezone
+from uuid import UUID, uuid4
+
+import pytest
+
+from reflector.db.daily_recording_requests import (
+    DailyRecordingRequest,
+    daily_recording_requests_controller,
+)
+from reflector.db.meetings import Meeting, meetings_controller
+from reflector.db.recordings import Recording, recordings_controller
+from reflector.db.rooms import Room, rooms_controller
+
+
+@pytest.mark.asyncio
+async def test_recording_request_flow_cloud(client):
+    """Test full cloud recording flow: start -> webhook -> match"""
+    # Create room and meeting
+    room = Room(id="test-room", name="Test Room", slug="test-room", user_id="test-user")
+    await rooms_controller.create(room)
+
+    meeting_id = f"meeting-{uuid4()}"
+    meeting = Meeting(
+        id=meeting_id,
+        room_name="test-room",
+        start_date=datetime.now(timezone.utc),
+        end_date=None,
+        recording_type="cloud",
+    )
+    await meetings_controller.create(meeting)
+
+    # Simulate recording start (what endpoint does)
+    recording_id = "rec-cloud-123"
+    instance_id = UUID("a1b2c3d4-e5f6-7890-abcd-ef1234567890")
+
+    request = DailyRecordingRequest(
+        recording_id=recording_id,
+        meeting_id=meeting_id,
+        instance_id=instance_id,
+        type="cloud",
+        requested_at=datetime.now(timezone.utc),
+    )
+    await daily_recording_requests_controller.create(request)
+
+    # Verify request exists
+    match = await daily_recording_requests_controller.find_by_recording_id(recording_id)
+    assert match is not None
+    assert match[0] == meeting_id
+    assert match[1] == "cloud"
+
+    # Simulate webhook/polling storing cloud recording
+    success = await meetings_controller.set_cloud_recording_if_missing(
+        meeting_id=meeting_id,
+        s3_key="s3://bucket/recording.mp4",
+        duration=120,
+    )
+    assert success is True
+
+    # Verify meeting updated
+    updated_meeting = await meetings_controller.get_by_id(meeting_id)
+    assert updated_meeting.daily_composed_video_s3_key == "s3://bucket/recording.mp4"
+    assert updated_meeting.daily_composed_video_duration == 120
+
+
+@pytest.mark.asyncio
+async def test_recording_request_flow_raw_tracks(client):
+    """Test full raw-tracks recording flow: start -> webhook/polling -> process"""
+    # Create room and meeting
+    room = Room(
+        id="test-room-2",
+        name="Test Room 2",
+        slug="test-room-2",
+        user_id="test-user",
+    )
+    await rooms_controller.create(room)
+
+    meeting_id = f"meeting-{uuid4()}"
+    meeting = Meeting(
+        id=meeting_id,
+        room_name="test-room-2",
+        start_date=datetime.now(timezone.utc),
+        end_date=None,
+        recording_type="raw-tracks",
+    )
+    await meetings_controller.create(meeting)
+
+    # Simulate recording start
+    recording_id = "rec-raw-456"
+    instance_id = UUID("b1c2d3e4-f5a6-7890-abcd-ef1234567890")
+
+    request = DailyRecordingRequest(
+        recording_id=recording_id,
+        meeting_id=meeting_id,
+        instance_id=instance_id,
+        type="raw-tracks",
+        requested_at=datetime.now(timezone.utc),
+    )
+    await daily_recording_requests_controller.create(request)
+
+    # Simulate webhook/polling discovering recording
+    match = await daily_recording_requests_controller.find_by_recording_id(recording_id)
+    assert match is not None
+    found_meeting_id, recording_type = match
+    assert found_meeting_id == meeting_id
+    assert recording_type == "raw-tracks"
+
+    # Create recording (what webhook/polling does)
+    created = await recordings_controller.try_create_with_meeting(
+        Recording(
+            id=recording_id,
+            bucket_name="test-bucket",
+            object_key="recordings/20260120/",
+            recorded_at=datetime.now(timezone.utc),
+            track_keys=["track1.webm", "track2.webm"],
+            meeting_id=meeting_id,
+            status="pending",
+        )
+    )
+    assert created is True
+
+    # Verify recording exists with meeting_id
+    recording = await recordings_controller.get_by_id(recording_id)
+    assert recording is not None
+    assert recording.meeting_id == meeting_id
+    assert recording.status == "pending"
+    assert len(recording.track_keys) == 2
+
+
+@pytest.mark.asyncio
+async def test_stop_restart_creates_multiple_requests(client):
+    """Test stop/restart creates multiple request rows with same instance_id"""
+    # Create room and meeting
+    room = Room(
+        id="test-room-3",
+        name="Test Room 3",
+        slug="test-room-3",
+        user_id="test-user",
+    )
+    await rooms_controller.create(room)
+
+    meeting_id = f"meeting-{uuid4()}"
+    meeting = Meeting(
+        id=meeting_id,
+        room_name="test-room-3",
+        start_date=datetime.now(timezone.utc),
+        end_date=None,
+        recording_type="cloud",
+    )
+    await meetings_controller.create(meeting)
+
+    instance_id = UUID("c1d2e3f4-a5b6-7890-abcd-ef1234567890")
+
+    # First recording
+    await daily_recording_requests_controller.create(
+        DailyRecordingRequest(
+            recording_id="rec-first",
+            meeting_id=meeting_id,
+            instance_id=instance_id,
+            type="cloud",
+            requested_at=datetime.now(timezone.utc),
+        )
+    )
+
+    # Stop, then restart (new recording_id, same instance_id)
+    await daily_recording_requests_controller.create(
+        DailyRecordingRequest(
+            recording_id="rec-second",  # DIFFERENT
+            meeting_id=meeting_id,
+            instance_id=instance_id,  # SAME
+            type="cloud",
+            requested_at=datetime.now(timezone.utc),
+        )
+    )
+
+    # Both exist
+    requests = await daily_recording_requests_controller.get_by_meeting_id(meeting_id)
+    assert len(requests) == 2
+    assert {r.recording_id for r in requests} == {"rec-first", "rec-second"}
+    assert all(r.instance_id == instance_id for r in requests)
+
+
+@pytest.mark.asyncio
+async def test_orphan_recording_no_request(client):
+    """Test orphan recording (no request found)"""
+    # Simulate polling discovering recording with no request
+    recording_id = "rec-orphan"
+
+    match = await daily_recording_requests_controller.find_by_recording_id(recording_id)
+    assert match is None  # No request
+
+    # Mark as orphan
+    created = await recordings_controller.create_orphan(
+        Recording(
+            id=recording_id,
+            bucket_name="test-bucket",
+            object_key="orphan-key",
+            recorded_at=datetime.now(timezone.utc),
+            meeting_id=None,
+            status="orphan",
+            track_keys=None,
+        )
+    )
+    assert created is True
+
+    # Verify orphan exists
+    recording = await recordings_controller.get_by_id(recording_id)
+    assert recording is not None
+    assert recording.status == "orphan"
+    assert recording.meeting_id is None
+
+    # Second poll - already exists
+    created_again = await recordings_controller.create_orphan(
+        Recording(
+            id=recording_id,
+            bucket_name="test-bucket",
+            object_key="orphan-key",
+            recorded_at=datetime.now(timezone.utc),
+            meeting_id=None,
+            status="orphan",
+            track_keys=None,
+        )
+    )
+    assert created_again is False  # Already exists
+
+
+@pytest.mark.asyncio
+async def test_concurrent_polling_deduplication(client):
+    """Test concurrent pollers only queue once"""
+    # Create room and meeting
+    room = Room(
+        id="test-room-4",
+        name="Test Room 4",
+        slug="test-room-4",
+        user_id="test-user",
+    )
+    await rooms_controller.create(room)
+
+    meeting_id = f"meeting-{uuid4()}"
+    meeting = Meeting(
+        id=meeting_id,
+        room_name="test-room-4",
+        start_date=datetime.now(timezone.utc),
+        end_date=None,
+        recording_type="raw-tracks",
+    )
+    await meetings_controller.create(meeting)
+
+    # Create request
+    recording_id = "rec-concurrent"
+    await daily_recording_requests_controller.create(
+        DailyRecordingRequest(
+            recording_id=recording_id,
+            meeting_id=meeting_id,
+            instance_id=UUID("d1e2f3a4-b5c6-7890-abcd-ef1234567890"),
+            type="raw-tracks",
+            requested_at=datetime.now(timezone.utc),
+        )
+    )
+
+    # Poller 1
+    created1 = await recordings_controller.try_create_with_meeting(
+        Recording(
+            id=recording_id,
+            bucket_name="test-bucket",
+            object_key="test-key",
+            recorded_at=datetime.now(timezone.utc),
+            meeting_id=meeting_id,
+            status="pending",
+            track_keys=["track1.webm"],
+        )
+    )
+    assert created1 is True  # First wins
+
+    # Poller 2 (concurrent)
+    created2 = await recordings_controller.try_create_with_meeting(
+        Recording(
+            id=recording_id,
+            bucket_name="test-bucket",
+            object_key="test-key",
+            recorded_at=datetime.now(timezone.utc),
+            meeting_id=meeting_id,
+            status="pending",
+            track_keys=["track1.webm"],
+        )
+    )
+    assert created2 is False  # Conflict, skip
+
+    # Only one recording exists
+    recording = await recordings_controller.get_by_id(recording_id)
+    assert recording is not None
+    assert recording.meeting_id == meeting_id

server/tests/test_storage.py

@@ -319,51 +319,3 @@ def test_aws_storage_constructor_rejects_mixed_auth():
             aws_secret_access_key="test-secret",
             aws_role_arn="arn:aws:iam::123456789012:role/test-role",
         )
-
-
-@pytest.mark.asyncio
-async def test_aws_storage_custom_endpoint_url():
-    """Test that custom endpoint_url configures path-style addressing and passes endpoint to client."""
-    storage = AwsStorage(
-        aws_bucket_name="reflector-media",
-        aws_region="garage",
-        aws_access_key_id="GKtest",
-        aws_secret_access_key="secret",
-        aws_endpoint_url="http://garage:3900",
-    )
-    assert storage._endpoint_url == "http://garage:3900"
-    assert storage.boto_config.s3["addressing_style"] == "path"
-    assert storage.base_url == "http://garage:3900/reflector-media/"
-    # retries config preserved (merge, not replace)
-    assert storage.boto_config.retries["max_attempts"] == 3
-
-    mock_client = AsyncMock()
-    mock_client.put_object = AsyncMock()
-    mock_client.__aenter__ = AsyncMock(return_value=mock_client)
-    mock_client.__aexit__ = AsyncMock(return_value=None)
-    mock_client.generate_presigned_url = AsyncMock(
-        return_value="http://garage:3900/reflector-media/test.txt"
-    )
-
-    with patch.object(
-        storage.session, "client", return_value=mock_client
-    ) as mock_session_client:
-        await storage.put_file("test.txt", b"data")
-        mock_session_client.assert_called_with(
-            "s3", config=storage.boto_config, endpoint_url="http://garage:3900"
-        )
-
-
-@pytest.mark.asyncio
-async def test_aws_storage_none_endpoint_url():
-    """Test that None endpoint preserves current AWS behavior."""
-    storage = AwsStorage(
-        aws_bucket_name="reflector-bucket",
-        aws_region="us-east-1",
-        aws_access_key_id="AKIAtest",
-        aws_secret_access_key="secret",
-    )
-    assert storage._endpoint_url is None
-    assert storage.base_url == "https://reflector-bucket.s3.amazonaws.com/"
-    # No s3 addressing_style override — boto_config should only have retries
-    assert not hasattr(storage.boto_config, "s3") or storage.boto_config.s3 is None

server/tests/test_time_based_meeting_matching.py

@@ -1,374 +0,0 @@
-"""
-Integration tests for time-based meeting-to-recording matching.
-
-Tests the critical path for matching Daily.co recordings to meetings when
-API doesn't return instanceId.
-"""
-
-from datetime import datetime, timedelta, timezone
-
-import pytest
-
-from reflector.db.meetings import meetings_controller
-from reflector.db.rooms import rooms_controller
-
-
-@pytest.fixture
-async def test_room():
-    """Create a test room for meetings."""
-    room = await rooms_controller.add(
-        name="test-room-time",
-        user_id="test-user-id",
-        zulip_auto_post=False,
-        zulip_stream="",
-        zulip_topic="",
-        is_locked=False,
-        room_mode="normal",
-        recording_type="cloud",
-        recording_trigger="automatic",
-        is_shared=False,
-        platform="daily",
-    )
-    return room
-
-
-@pytest.fixture
-def base_time():
-    """Fixed timestamp for deterministic tests."""
-    return datetime(2026, 1, 14, 9, 0, 0, tzinfo=timezone.utc)
-
-
-class TestTimeBasedMatching:
-    """Test get_by_room_name_and_time() matching logic."""
-
-    async def test_exact_time_match(self, test_room, base_time):
-        """Recording timestamp exactly matches meeting start_date."""
-        meeting = await meetings_controller.create(
-            id="meeting-exact",
-            room_name="daily-test-20260114090000",
-            room_url="https://example.daily.co/test",
-            host_room_url="https://example.daily.co/test?t=host",
-            start_date=base_time,
-            end_date=base_time + timedelta(hours=1),
-            room=test_room,
-        )
-
-        result = await meetings_controller.get_by_room_name_and_time(
-            room_name="daily-test-20260114090000",
-            recording_start=base_time,
-            time_window_hours=168,
-        )
-
-        assert result is not None
-        assert result.id == meeting.id
-
-    async def test_recording_slightly_after_meeting_start(self, test_room, base_time):
-        """Recording started 1 minute after meeting (participants joined late)."""
-        meeting = await meetings_controller.create(
-            id="meeting-late",
-            room_name="daily-test-20260114090100",
-            room_url="https://example.daily.co/test",
-            host_room_url="https://example.daily.co/test?t=host",
-            start_date=base_time,
-            end_date=base_time + timedelta(hours=1),
-            room=test_room,
-        )
-
-        recording_start = base_time + timedelta(minutes=1)
-
-        result = await meetings_controller.get_by_room_name_and_time(
-            room_name="daily-test-20260114090100",
-            recording_start=recording_start,
-            time_window_hours=168,
-        )
-
-        assert result is not None
-        assert result.id == meeting.id
-
-    async def test_duplicate_room_names_picks_closest(self, test_room, base_time):
-        """
-        Two meetings with same room_name (duplicate/race condition).
-        Should pick closest by timestamp.
-        """
-        meeting1 = await meetings_controller.create(
-            id="meeting-1-first",
-            room_name="daily-duplicate-room",
-            room_url="https://example.daily.co/test",
-            host_room_url="https://example.daily.co/test?t=host",
-            start_date=base_time,
-            end_date=base_time + timedelta(hours=1),
-            room=test_room,
-        )
-
-        meeting2 = await meetings_controller.create(
-            id="meeting-2-second",
-            room_name="daily-duplicate-room",  # Same room_name!
-            room_url="https://example.daily.co/test",
-            host_room_url="https://example.daily.co/test?t=host",
-            start_date=base_time + timedelta(seconds=0.99),  # 0.99s later
-            end_date=base_time + timedelta(hours=1),
-            room=test_room,
-        )
-
-        # Recording started 0.5s after meeting1
-        # Distance: meeting1 = 0.5s, meeting2 = 0.49s → meeting2 is closer
-        recording_start = base_time + timedelta(seconds=0.5)
-
-        result = await meetings_controller.get_by_room_name_and_time(
-            room_name="daily-duplicate-room",
-            recording_start=recording_start,
-            time_window_hours=168,
-        )
-
-        assert result is not None
-        assert result.id == meeting2.id  # meeting2 is closer (0.49s vs 0.5s)
-
-    async def test_outside_time_window_returns_none(self, test_room, base_time):
-        """Recording outside 1-week window returns None."""
-        await meetings_controller.create(
-            id="meeting-old",
-            room_name="daily-test-old",
-            room_url="https://example.daily.co/test",
-            host_room_url="https://example.daily.co/test?t=host",
-            start_date=base_time,
-            end_date=base_time + timedelta(hours=1),
-            room=test_room,
-        )
-
-        # Recording 8 days later (outside 7-day window)
-        recording_start = base_time + timedelta(days=8)
-
-        result = await meetings_controller.get_by_room_name_and_time(
-            room_name="daily-test-old",
-            recording_start=recording_start,
-            time_window_hours=168,
-        )
-
-        assert result is None
-
-    async def test_tie_breaker_deterministic(self, test_room, base_time):
-        """When time delta identical, tie-breaker by meeting.id is deterministic."""
-        meeting_z = await meetings_controller.create(
-            id="zzz-last-uuid",
-            room_name="daily-test-tie",
-            room_url="https://example.daily.co/test",
-            host_room_url="https://example.daily.co/test?t=host",
-            start_date=base_time,
-            end_date=base_time + timedelta(hours=1),
-            room=test_room,
-        )
-
-        meeting_a = await meetings_controller.create(
-            id="aaa-first-uuid",
-            room_name="daily-test-tie",
-            room_url="https://example.daily.co/test",
-            host_room_url="https://example.daily.co/test?t=host",
-            start_date=base_time,  # Exact same start_date
-            end_date=base_time + timedelta(hours=1),
-            room=test_room,
-        )
-
-        result = await meetings_controller.get_by_room_name_and_time(
-            room_name="daily-test-tie",
-            recording_start=base_time,
-            time_window_hours=168,
-        )
-
-        assert result is not None
-        # Tie-breaker: lexicographically first UUID
-        assert result.id == "aaa-first-uuid"
-
-    async def test_timezone_naive_datetime_raises(self, test_room, base_time):
-        """Timezone-naive datetime raises ValueError."""
-        await meetings_controller.create(
-            id="meeting-tz",
-            room_name="daily-test-tz",
-            room_url="https://example.daily.co/test",
-            host_room_url="https://example.daily.co/test?t=host",
-            start_date=base_time,
-            end_date=base_time + timedelta(hours=1),
-            room=test_room,
-        )
-
-        # Naive datetime (no timezone)
-        naive_dt = datetime(2026, 1, 14, 9, 0, 0)
-
-        with pytest.raises(ValueError, match="timezone-aware"):
-            await meetings_controller.get_by_room_name_and_time(
-                room_name="daily-test-tz",
-                recording_start=naive_dt,
-                time_window_hours=168,
-            )
-
-    async def test_one_week_boundary_after_included(self, test_room, base_time):
-        """Meeting 1-week AFTER recording is included (window_end boundary)."""
-        meeting_time = base_time + timedelta(hours=168)
-
-        await meetings_controller.create(
-            id="meeting-boundary-after",
-            room_name="daily-test-boundary-after",
-            room_url="https://example.daily.co/test",
-            host_room_url="https://example.daily.co/test?t=host",
-            start_date=meeting_time,
-            end_date=meeting_time + timedelta(hours=1),
-            room=test_room,
-        )
-
-        result = await meetings_controller.get_by_room_name_and_time(
-            room_name="daily-test-boundary-after",
-            recording_start=base_time,
-            time_window_hours=168,
-        )
-
-        assert result is not None
-        assert result.id == "meeting-boundary-after"
-
-    async def test_one_week_boundary_before_included(self, test_room, base_time):
-        """Meeting 1-week BEFORE recording is included (window_start boundary)."""
-        meeting_time = base_time - timedelta(hours=168)
-
-        await meetings_controller.create(
-            id="meeting-boundary-before",
-            room_name="daily-test-boundary-before",
-            room_url="https://example.daily.co/test",
-            host_room_url="https://example.daily.co/test?t=host",
-            start_date=meeting_time,
-            end_date=meeting_time + timedelta(hours=1),
-            room=test_room,
-        )
-
-        result = await meetings_controller.get_by_room_name_and_time(
-            room_name="daily-test-boundary-before",
-            recording_start=base_time,
-            time_window_hours=168,
-        )
-
-        assert result is not None
-        assert result.id == "meeting-boundary-before"
-
-    async def test_recording_before_meeting_start(self, test_room, base_time):
-        """Recording started before meeting (clock skew or early join)."""
-        await meetings_controller.create(
-            id="meeting-early",
-            room_name="daily-test-early",
-            room_url="https://example.daily.co/test",
-            host_room_url="https://example.daily.co/test?t=host",
-            start_date=base_time,
-            end_date=base_time + timedelta(hours=1),
-            room=test_room,
-        )
-
-        recording_start = base_time - timedelta(minutes=2)
-
-        result = await meetings_controller.get_by_room_name_and_time(
-            room_name="daily-test-early",
-            recording_start=recording_start,
-            time_window_hours=168,
-        )
-
-        assert result is not None
-        assert result.id == "meeting-early"
-
-    async def test_mixed_inside_outside_window(self, test_room, base_time):
-        """Multiple meetings, only one inside window - returns the inside one."""
-        await meetings_controller.create(
-            id="meeting-old",
-            room_name="daily-test-mixed",
-            room_url="https://example.daily.co/test",
-            host_room_url="https://example.daily.co/test?t=host",
-            start_date=base_time - timedelta(days=10),
-            end_date=base_time - timedelta(days=10, hours=-1),
-            room=test_room,
-        )
-
-        await meetings_controller.create(
-            id="meeting-inside",
-            room_name="daily-test-mixed",
-            room_url="https://example.daily.co/test",
-            host_room_url="https://example.daily.co/test?t=host",
-            start_date=base_time - timedelta(days=2),
-            end_date=base_time - timedelta(days=2, hours=-1),
-            room=test_room,
-        )
-
-        await meetings_controller.create(
-            id="meeting-future",
-            room_name="daily-test-mixed",
-            room_url="https://example.daily.co/test",
-            host_room_url="https://example.daily.co/test?t=host",
-            start_date=base_time + timedelta(days=10),
-            end_date=base_time + timedelta(days=10, hours=1),
-            room=test_room,
-        )
-
-        result = await meetings_controller.get_by_room_name_and_time(
-            room_name="daily-test-mixed",
-            recording_start=base_time,
-            time_window_hours=168,
-        )
-
-        assert result is not None
-        assert result.id == "meeting-inside"
-
-
-class TestAtomicCloudRecordingUpdate:
-    """Test atomic update prevents race conditions."""
-
-    async def test_first_update_succeeds(self, test_room, base_time):
-        """First call to set_cloud_recording_if_missing succeeds."""
-        meeting = await meetings_controller.create(
-            id="meeting-atomic-1",
-            room_name="daily-test-atomic",
-            room_url="https://example.daily.co/test",
-            host_room_url="https://example.daily.co/test?t=host",
-            start_date=base_time,
-            end_date=base_time + timedelta(hours=1),
-            room=test_room,
-        )
-
-        success = await meetings_controller.set_cloud_recording_if_missing(
-            meeting_id=meeting.id,
-            s3_key="first-s3-key",
-            duration=100,
-        )
-
-        assert success is True
-
-        updated = await meetings_controller.get_by_id(meeting.id)
-        assert updated.daily_composed_video_s3_key == "first-s3-key"
-        assert updated.daily_composed_video_duration == 100
-
-    async def test_second_update_fails_atomically(self, test_room, base_time):
-        """Second call to update same meeting doesn't overwrite (atomic check)."""
-        meeting = await meetings_controller.create(
-            id="meeting-atomic-2",
-            room_name="daily-test-atomic2",
-            room_url="https://example.daily.co/test",
-            host_room_url="https://example.daily.co/test?t=host",
-            start_date=base_time,
-            end_date=base_time + timedelta(hours=1),
-            room=test_room,
-        )
-
-        success1 = await meetings_controller.set_cloud_recording_if_missing(
-            meeting_id=meeting.id,
-            s3_key="first-s3-key",
-            duration=100,
-        )
-
-        assert success1 is True
-
-        after_first = await meetings_controller.get_by_id(meeting.id)
-        assert after_first.daily_composed_video_s3_key == "first-s3-key"
-
-        success2 = await meetings_controller.set_cloud_recording_if_missing(
-            meeting_id=meeting.id,
-            s3_key="bucket/path/should-not-overwrite",
-            duration=200,
-        )
-
-        assert success2 is False
-
-        final = await meetings_controller.get_by_id(meeting.id)
-        assert final.daily_composed_video_s3_key == "first-s3-key"
-        assert final.daily_composed_video_duration == 100

server/tests/test_transcripts_process.py

@@ -1,6 +1,6 @@
 import asyncio
 import time
-from unittest.mock import AsyncMock, patch
+from unittest.mock import patch
 
 import pytest
 from httpx import ASGITransport, AsyncClient
@@ -142,17 +142,17 @@ async def test_whereby_recording_uses_file_pipeline(client):
             "reflector.services.transcript_process.task_pipeline_file_process"
         ) as mock_file_pipeline,
         patch(
-            "reflector.services.transcript_process.HatchetClientManager"
-        ) as mock_hatchet,
+            "reflector.services.transcript_process.task_pipeline_multitrack_process"
+        ) as mock_multitrack_pipeline,
     ):
         response = await client.post(f"/transcripts/{transcript.id}/process")
 
         assert response.status_code == 200
         assert response.json()["status"] == "ok"
 
-        # Whereby recordings should use file pipeline, not Hatchet
+        # Whereby recordings should use file pipeline
         mock_file_pipeline.delay.assert_called_once_with(transcript_id=transcript.id)
-        mock_hatchet.start_workflow.assert_not_called()
+        mock_multitrack_pipeline.delay.assert_not_called()
 
 
 @pytest.mark.usefixtures("setup_database")
@@ -177,6 +177,8 @@ async def test_dailyco_recording_uses_multitrack_pipeline(client):
         recording_trigger="automatic-2nd-participant",
         is_shared=False,
     )
+    # Force Celery backend for test
+    await rooms_controller.update(room, {"use_celery": True})
 
     transcript = await transcripts_controller.add(
         "",
@@ -211,23 +213,18 @@ async def test_dailyco_recording_uses_multitrack_pipeline(client):
             "reflector.services.transcript_process.task_pipeline_file_process"
         ) as mock_file_pipeline,
         patch(
-            "reflector.services.transcript_process.HatchetClientManager"
-        ) as mock_hatchet,
+            "reflector.services.transcript_process.task_pipeline_multitrack_process"
+        ) as mock_multitrack_pipeline,
     ):
-        mock_hatchet.start_workflow = AsyncMock(return_value="test-workflow-id")
-
         response = await client.post(f"/transcripts/{transcript.id}/process")
 
         assert response.status_code == 200
         assert response.json()["status"] == "ok"
 
-        # Daily.co multitrack recordings should use Hatchet workflow
-        mock_hatchet.start_workflow.assert_called_once()
-        call_kwargs = mock_hatchet.start_workflow.call_args.kwargs
-        assert call_kwargs["workflow_name"] == "DiarizationPipeline"
-        assert call_kwargs["input_data"]["transcript_id"] == transcript.id
-        assert call_kwargs["input_data"]["bucket_name"] == "daily-bucket"
-        assert call_kwargs["input_data"]["tracks"] == [
-            {"s3_key": k} for k in track_keys
-        ]
+        # Daily.co multitrack recordings should use multitrack pipeline
+        mock_multitrack_pipeline.delay.assert_called_once_with(
+            transcript_id=transcript.id,
+            bucket_name="daily-bucket",
+            track_keys=track_keys,
+        )
         mock_file_pipeline.delay.assert_not_called()

server/uv.lock

@@ -235,6 +235,12 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/78/b6/6307fbef88d9b5ee7421e68d78a9f162e0da4900bc5f5793f6d3d0e34fb8/annotated_types-0.7.0-py3-none-any.whl", hash = "sha256:1f02e8b43a8fbbc3f3e0d4f0f4bfc8131bcb4eebe8849b8e5c773f3a1c582a53", size = 13643 },
 ]
 
+[[package]]
+name = "antlr4-python3-runtime"
+version = "4.9.3"
+source = { registry = "https://pypi.org/simple" }
+sdist = { url = "https://files.pythonhosted.org/packages/3e/38/7859ff46355f76f8d19459005ca000b6e7012f2f1ca597746cbcd1fbfe5e/antlr4-python3-runtime-4.9.3.tar.gz", hash = "sha256:f224469b4168294902bb1efa80a8bf7855f24c99aef99cbefc1bcd3cce77881b", size = 117034 }
+
 [[package]]
 name = "anyio"
 version = "4.9.0"
@@ -261,6 +267,21 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/2f/f5/c36551e93acba41a59939ae6a0fb77ddb3f2e8e8caa716410c65f7341f72/asgi_lifespan-2.1.0-py3-none-any.whl", hash = "sha256:ed840706680e28428c01e14afb3875d7d76d3206f3d5b2f2294e059b5c23804f", size = 10895 },
 ]
 
+[[package]]
+name = "asteroid-filterbanks"
+version = "0.4.0"
+source = { registry = "https://pypi.org/simple" }
+dependencies = [
+    { name = "numpy" },
+    { name = "torch", version = "2.8.0", source = { registry = "https://download.pytorch.org/whl/cpu" }, marker = "sys_platform == 'darwin'" },
+    { name = "torch", version = "2.8.0+cpu", source = { registry = "https://download.pytorch.org/whl/cpu" }, marker = "sys_platform != 'darwin'" },
+    { name = "typing-extensions" },
+]
+sdist = { url = "https://files.pythonhosted.org/packages/90/fa/5c2be1f96dc179f83cdd3bb267edbd1f47d08f756785c016d5c2163901a7/asteroid-filterbanks-0.4.0.tar.gz", hash = "sha256:415f89d1dcf2b13b35f03f7a9370968ac4e6fa6800633c522dac992b283409b9", size = 24599 }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/c5/7c/83ff6046176a675e6a1e8aeefed8892cd97fe7c46af93cc540d1b24b8323/asteroid_filterbanks-0.4.0-py3-none-any.whl", hash = "sha256:4932ac8b6acc6e08fb87cbe8ece84215b5a74eee284fe83acf3540a72a02eaf5", size = 29912 },
+]
+
 [[package]]
 name = "async-timeout"
 version = "5.0.1"
@@ -582,6 +603,56 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/a7/06/3d6badcf13db419e25b07041d9c7b4a2c331d3f4e7134445ec5df57714cd/coloredlogs-15.0.1-py2.py3-none-any.whl", hash = "sha256:612ee75c546f53e92e70049c9dbfcc18c935a2b9a53b66085ce9ef6a6e5c0934", size = 46018 },
 ]
 
+[[package]]
+name = "colorlog"
+version = "6.9.0"
+source = { registry = "https://pypi.org/simple" }
+dependencies = [
+    { name = "colorama", marker = "sys_platform == 'win32'" },
+]
+sdist = { url = "https://files.pythonhosted.org/packages/d3/7a/359f4d5df2353f26172b3cc39ea32daa39af8de522205f512f458923e677/colorlog-6.9.0.tar.gz", hash = "sha256:bfba54a1b93b94f54e1f4fe48395725a3d92fd2a4af702f6bd70946bdc0c6ac2", size = 16624 }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/e3/51/9b208e85196941db2f0654ad0357ca6388ab3ed67efdbfc799f35d1f83aa/colorlog-6.9.0-py3-none-any.whl", hash = "sha256:5906e71acd67cb07a71e779c47c4bcb45fb8c2993eebe9e5adcd6a6f1b283eff", size = 11424 },
+]
+
+[[package]]
+name = "contourpy"
+version = "1.3.3"
+source = { registry = "https://pypi.org/simple" }
+dependencies = [
+    { name = "numpy" },
+]
+sdist = { url = "https://files.pythonhosted.org/packages/58/01/1253e6698a07380cd31a736d248a3f2a50a7c88779a1813da27503cadc2a/contourpy-1.3.3.tar.gz", hash = "sha256:083e12155b210502d0bca491432bb04d56dc3432f95a979b429f2848c3dbe880", size = 13466174 }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/91/2e/c4390a31919d8a78b90e8ecf87cd4b4c4f05a5b48d05ec17db8e5404c6f4/contourpy-1.3.3-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:709a48ef9a690e1343202916450bc48b9e51c049b089c7f79a267b46cffcdaa1", size = 288773 },
+    { url = "https://files.pythonhosted.org/packages/0d/44/c4b0b6095fef4dc9c420e041799591e3b63e9619e3044f7f4f6c21c0ab24/contourpy-1.3.3-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:23416f38bfd74d5d28ab8429cc4d63fa67d5068bd711a85edb1c3fb0c3e2f381", size = 270149 },
+    { url = "https://files.pythonhosted.org/packages/30/2e/dd4ced42fefac8470661d7cb7e264808425e6c5d56d175291e93890cce09/contourpy-1.3.3-cp311-cp311-manylinux_2_26_aarch64.manylinux_2_28_aarch64.whl", hash = "sha256:929ddf8c4c7f348e4c0a5a3a714b5c8542ffaa8c22954862a46ca1813b667ee7", size = 329222 },
+    { url = "https://files.pythonhosted.org/packages/f2/74/cc6ec2548e3d276c71389ea4802a774b7aa3558223b7bade3f25787fafc2/contourpy-1.3.3-cp311-cp311-manylinux_2_26_ppc64le.manylinux_2_28_ppc64le.whl", hash = "sha256:9e999574eddae35f1312c2b4b717b7885d4edd6cb46700e04f7f02db454e67c1", size = 377234 },
+    { url = "https://files.pythonhosted.org/packages/03/b3/64ef723029f917410f75c09da54254c5f9ea90ef89b143ccadb09df14c15/contourpy-1.3.3-cp311-cp311-manylinux_2_26_s390x.manylinux_2_28_s390x.whl", hash = "sha256:0bf67e0e3f482cb69779dd3061b534eb35ac9b17f163d851e2a547d56dba0a3a", size = 380555 },
+    { url = "https://files.pythonhosted.org/packages/5f/4b/6157f24ca425b89fe2eb7e7be642375711ab671135be21e6faa100f7448c/contourpy-1.3.3-cp311-cp311-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl", hash = "sha256:51e79c1f7470158e838808d4a996fa9bac72c498e93d8ebe5119bc1e6becb0db", size = 355238 },
+    { url = "https://files.pythonhosted.org/packages/98/56/f914f0dd678480708a04cfd2206e7c382533249bc5001eb9f58aa693e200/contourpy-1.3.3-cp311-cp311-musllinux_1_2_aarch64.whl", hash = "sha256:598c3aaece21c503615fd59c92a3598b428b2f01bfb4b8ca9c4edeecc2438620", size = 1326218 },
+    { url = "https://files.pythonhosted.org/packages/fb/d7/4a972334a0c971acd5172389671113ae82aa7527073980c38d5868ff1161/contourpy-1.3.3-cp311-cp311-musllinux_1_2_x86_64.whl", hash = "sha256:322ab1c99b008dad206d406bb61d014cf0174df491ae9d9d0fac6a6fda4f977f", size = 1392867 },
+    { url = "https://files.pythonhosted.org/packages/75/3e/f2cc6cd56dc8cff46b1a56232eabc6feea52720083ea71ab15523daab796/contourpy-1.3.3-cp311-cp311-win32.whl", hash = "sha256:fd907ae12cd483cd83e414b12941c632a969171bf90fc937d0c9f268a31cafff", size = 183677 },
+    { url = "https://files.pythonhosted.org/packages/98/4b/9bd370b004b5c9d8045c6c33cf65bae018b27aca550a3f657cdc99acdbd8/contourpy-1.3.3-cp311-cp311-win_amd64.whl", hash = "sha256:3519428f6be58431c56581f1694ba8e50626f2dd550af225f82fb5f5814d2a42", size = 225234 },
+    { url = "https://files.pythonhosted.org/packages/d9/b6/71771e02c2e004450c12b1120a5f488cad2e4d5b590b1af8bad060360fe4/contourpy-1.3.3-cp311-cp311-win_arm64.whl", hash = "sha256:15ff10bfada4bf92ec8b31c62bf7c1834c244019b4a33095a68000d7075df470", size = 193123 },
+    { url = "https://files.pythonhosted.org/packages/be/45/adfee365d9ea3d853550b2e735f9d66366701c65db7855cd07621732ccfc/contourpy-1.3.3-cp312-cp312-macosx_10_13_x86_64.whl", hash = "sha256:b08a32ea2f8e42cf1d4be3169a98dd4be32bafe4f22b6c4cb4ba810fa9e5d2cb", size = 293419 },
+    { url = "https://files.pythonhosted.org/packages/53/3e/405b59cfa13021a56bba395a6b3aca8cec012b45bf177b0eaf7a202cde2c/contourpy-1.3.3-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:556dba8fb6f5d8742f2923fe9457dbdd51e1049c4a43fd3986a0b14a1d815fc6", size = 273979 },
+    { url = "https://files.pythonhosted.org/packages/d4/1c/a12359b9b2ca3a845e8f7f9ac08bdf776114eb931392fcad91743e2ea17b/contourpy-1.3.3-cp312-cp312-manylinux_2_26_aarch64.manylinux_2_28_aarch64.whl", hash = "sha256:92d9abc807cf7d0e047b95ca5d957cf4792fcd04e920ca70d48add15c1a90ea7", size = 332653 },
+    { url = "https://files.pythonhosted.org/packages/63/12/897aeebfb475b7748ea67b61e045accdfcf0d971f8a588b67108ed7f5512/contourpy-1.3.3-cp312-cp312-manylinux_2_26_ppc64le.manylinux_2_28_ppc64le.whl", hash = "sha256:b2e8faa0ed68cb29af51edd8e24798bb661eac3bd9f65420c1887b6ca89987c8", size = 379536 },
+    { url = "https://files.pythonhosted.org/packages/43/8a/a8c584b82deb248930ce069e71576fc09bd7174bbd35183b7943fb1064fd/contourpy-1.3.3-cp312-cp312-manylinux_2_26_s390x.manylinux_2_28_s390x.whl", hash = "sha256:626d60935cf668e70a5ce6ff184fd713e9683fb458898e4249b63be9e28286ea", size = 384397 },
+    { url = "https://files.pythonhosted.org/packages/cc/8f/ec6289987824b29529d0dfda0d74a07cec60e54b9c92f3c9da4c0ac732de/contourpy-1.3.3-cp312-cp312-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl", hash = "sha256:4d00e655fcef08aba35ec9610536bfe90267d7ab5ba944f7032549c55a146da1", size = 362601 },
+    { url = "https://files.pythonhosted.org/packages/05/0a/a3fe3be3ee2dceb3e615ebb4df97ae6f3828aa915d3e10549ce016302bd1/contourpy-1.3.3-cp312-cp312-musllinux_1_2_aarch64.whl", hash = "sha256:451e71b5a7d597379ef572de31eeb909a87246974d960049a9848c3bc6c41bf7", size = 1331288 },
+    { url = "https://files.pythonhosted.org/packages/33/1d/acad9bd4e97f13f3e2b18a3977fe1b4a37ecf3d38d815333980c6c72e963/contourpy-1.3.3-cp312-cp312-musllinux_1_2_x86_64.whl", hash = "sha256:459c1f020cd59fcfe6650180678a9993932d80d44ccde1fa1868977438f0b411", size = 1403386 },
+    { url = "https://files.pythonhosted.org/packages/cf/8f/5847f44a7fddf859704217a99a23a4f6417b10e5ab1256a179264561540e/contourpy-1.3.3-cp312-cp312-win32.whl", hash = "sha256:023b44101dfe49d7d53932be418477dba359649246075c996866106da069af69", size = 185018 },
+    { url = "https://files.pythonhosted.org/packages/19/e8/6026ed58a64563186a9ee3f29f41261fd1828f527dd93d33b60feca63352/contourpy-1.3.3-cp312-cp312-win_amd64.whl", hash = "sha256:8153b8bfc11e1e4d75bcb0bff1db232f9e10b274e0929de9d608027e0d34ff8b", size = 226567 },
+    { url = "https://files.pythonhosted.org/packages/d1/e2/f05240d2c39a1ed228d8328a78b6f44cd695f7ef47beb3e684cf93604f86/contourpy-1.3.3-cp312-cp312-win_arm64.whl", hash = "sha256:07ce5ed73ecdc4a03ffe3e1b3e3c1166db35ae7584be76f65dbbe28a7791b0cc", size = 193655 },
+    { url = "https://files.pythonhosted.org/packages/a5/29/8dcfe16f0107943fa92388c23f6e05cff0ba58058c4c95b00280d4c75a14/contourpy-1.3.3-pp311-pypy311_pp73-macosx_10_15_x86_64.whl", hash = "sha256:cd5dfcaeb10f7b7f9dc8941717c6c2ade08f587be2226222c12b25f0483ed497", size = 278809 },
+    { url = "https://files.pythonhosted.org/packages/85/a9/8b37ef4f7dafeb335daee3c8254645ef5725be4d9c6aa70b50ec46ef2f7e/contourpy-1.3.3-pp311-pypy311_pp73-macosx_11_0_arm64.whl", hash = "sha256:0c1fc238306b35f246d61a1d416a627348b5cf0648648a031e14bb8705fcdfe8", size = 261593 },
+    { url = "https://files.pythonhosted.org/packages/0a/59/ebfb8c677c75605cc27f7122c90313fd2f375ff3c8d19a1694bda74aaa63/contourpy-1.3.3-pp311-pypy311_pp73-manylinux_2_26_aarch64.manylinux_2_28_aarch64.whl", hash = "sha256:70f9aad7de812d6541d29d2bbf8feb22ff7e1c299523db288004e3157ff4674e", size = 302202 },
+    { url = "https://files.pythonhosted.org/packages/3c/37/21972a15834d90bfbfb009b9d004779bd5a07a0ec0234e5ba8f64d5736f4/contourpy-1.3.3-pp311-pypy311_pp73-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl", hash = "sha256:5ed3657edf08512fc3fe81b510e35c2012fbd3081d2e26160f27ca28affec989", size = 329207 },
+    { url = "https://files.pythonhosted.org/packages/0c/58/bd257695f39d05594ca4ad60df5bcb7e32247f9951fd09a9b8edb82d1daa/contourpy-1.3.3-pp311-pypy311_pp73-win_amd64.whl", hash = "sha256:3d1a3799d62d45c18bafd41c5fa05120b96a28079f2393af559b843d1a966a77", size = 225315 },
+]
+
 [[package]]
 name = "coverage"
 version = "7.9.2"
@@ -682,6 +753,15 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/ec/4c/0ecd260233290bee4b2facec4d8e755e57d8781d68f276e1248433993c9f/ctranslate2-4.6.0-cp312-cp312-win_amd64.whl", hash = "sha256:511cdf810a5bf6a2cec735799e5cd47966e63f8f7688fdee1b97fed621abda00", size = 19470040 },
 ]
 
+[[package]]
+name = "cycler"
+version = "0.12.1"
+source = { registry = "https://pypi.org/simple" }
+sdist = { url = "https://files.pythonhosted.org/packages/a9/95/a3dbbb5028f35eafb79008e7522a75244477d2838f38cbb722248dabc2a8/cycler-0.12.1.tar.gz", hash = "sha256:88bb128f02ba341da8ef447245a9e138fae777f6a23943da4540077d3601eb1c", size = 7615 }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/e7/05/c19819d5e3d95294a6f5947fb9b9629efb316b96de511b418c53d245aae6/cycler-0.12.1-py3-none-any.whl", hash = "sha256:85cef7cff222d8644161529808465972e51340599459b8ac3ccbac5a854e0d30", size = 8321 },
+]
+
 [[package]]
 name = "databases"
 version = "0.8.0"
@@ -794,6 +874,12 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/e3/26/57c6fb270950d476074c087527a558ccb6f4436657314bfb6cdf484114c4/docker-7.1.0-py3-none-any.whl", hash = "sha256:c96b93b7f0a746f9e77d325bcfb87422a3d8bd4f03136ae8a85b37f1898d5fc0", size = 147774 },
 ]
 
+[[package]]
+name = "docopt"
+version = "0.6.2"
+source = { registry = "https://pypi.org/simple" }
+sdist = { url = "https://files.pythonhosted.org/packages/a2/55/8f8cab2afd404cf578136ef2cc5dfb50baa1761b68c9da1fb1e4eed343c9/docopt-0.6.2.tar.gz", hash = "sha256:49b3a825280bd66b3aa83585ef59c4a8c82f2c8a522dbe754a8bc8d08c85c491", size = 25901 }
+
 [[package]]
 name = "ecdsa"
 version = "0.19.1"
@@ -806,6 +892,15 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/cb/a3/460c57f094a4a165c84a1341c373b0a4f5ec6ac244b998d5021aade89b77/ecdsa-0.19.1-py2.py3-none-any.whl", hash = "sha256:30638e27cf77b7e15c4c4cc1973720149e1033827cfd00661ca5c8cc0cdb24c3", size = 150607 },
 ]
 
+[[package]]
+name = "einops"
+version = "0.8.1"
+source = { registry = "https://pypi.org/simple" }
+sdist = { url = "https://files.pythonhosted.org/packages/e5/81/df4fbe24dff8ba3934af99044188e20a98ed441ad17a274539b74e82e126/einops-0.8.1.tar.gz", hash = "sha256:de5d960a7a761225532e0f1959e5315ebeafc0cd43394732f103ca44b9837e84", size = 54805 }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/87/62/9773de14fe6c45c23649e98b83231fffd7b9892b6cf863251dc2afa73643/einops-0.8.1-py3-none-any.whl", hash = "sha256:919387eb55330f5757c6bea9165c5ff5cfe63a642682ea788a6d472576d81737", size = 64359 },
+]
+
 [[package]]
 name = "email-validator"
 version = "2.2.0"
@@ -939,6 +1034,31 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/b8/25/155f9f080d5e4bc0082edfda032ea2bc2b8fab3f4d25d46c1e9dd22a1a89/flatbuffers-25.2.10-py2.py3-none-any.whl", hash = "sha256:ebba5f4d5ea615af3f7fd70fc310636fbb2bbd1f566ac0a23d98dd412de50051", size = 30953 },
 ]
 
+[[package]]
+name = "fonttools"
+version = "4.59.2"
+source = { registry = "https://pypi.org/simple" }
+sdist = { url = "https://files.pythonhosted.org/packages/0d/a5/fba25f9fbdab96e26dedcaeeba125e5f05a09043bf888e0305326e55685b/fonttools-4.59.2.tar.gz", hash = "sha256:e72c0749b06113f50bcb80332364c6be83a9582d6e3db3fe0b280f996dc2ef22", size = 3540889 }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/f8/53/742fcd750ae0bdc74de4c0ff923111199cc2f90a4ee87aaddad505b6f477/fonttools-4.59.2-cp311-cp311-macosx_10_9_universal2.whl", hash = "sha256:511946e8d7ea5c0d6c7a53c4cb3ee48eda9ab9797cd9bf5d95829a398400354f", size = 2774961 },
+    { url = "https://files.pythonhosted.org/packages/57/2a/976f5f9fa3b4dd911dc58d07358467bec20e813d933bc5d3db1a955dd456/fonttools-4.59.2-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:8e5e2682cf7be766d84f462ba8828d01e00c8751a8e8e7ce12d7784ccb69a30d", size = 2344690 },
+    { url = "https://files.pythonhosted.org/packages/c1/8f/b7eefc274fcf370911e292e95565c8253b0b87c82a53919ab3c795a4f50e/fonttools-4.59.2-cp311-cp311-manylinux2014_aarch64.manylinux_2_17_aarch64.manylinux_2_28_aarch64.whl", hash = "sha256:5729e12a982dba3eeae650de48b06f3b9ddb51e9aee2fcaf195b7d09a96250e2", size = 5026910 },
+    { url = "https://files.pythonhosted.org/packages/69/95/864726eaa8f9d4e053d0c462e64d5830ec7c599cbdf1db9e40f25ca3972e/fonttools-4.59.2-cp311-cp311-manylinux2014_x86_64.manylinux_2_17_x86_64.whl", hash = "sha256:c52694eae5d652361d59ecdb5a2246bff7cff13b6367a12da8499e9df56d148d", size = 4971031 },
+    { url = "https://files.pythonhosted.org/packages/24/4c/b8c4735ebdea20696277c70c79e0de615dbe477834e5a7c2569aa1db4033/fonttools-4.59.2-cp311-cp311-musllinux_1_2_aarch64.whl", hash = "sha256:f1f1bbc23ba1312bd8959896f46f667753b90216852d2a8cfa2d07e0cb234144", size = 5006112 },
+    { url = "https://files.pythonhosted.org/packages/3b/23/f9ea29c292aa2fc1ea381b2e5621ac436d5e3e0a5dee24ffe5404e58eae8/fonttools-4.59.2-cp311-cp311-musllinux_1_2_x86_64.whl", hash = "sha256:1a1bfe5378962825dabe741720885e8b9ae9745ec7ecc4a5ec1f1ce59a6062bf", size = 5117671 },
+    { url = "https://files.pythonhosted.org/packages/ba/07/cfea304c555bf06e86071ff2a3916bc90f7c07ec85b23bab758d4908c33d/fonttools-4.59.2-cp311-cp311-win32.whl", hash = "sha256:e937790f3c2c18a1cbc7da101550a84319eb48023a715914477d2e7faeaba570", size = 2218157 },
+    { url = "https://files.pythonhosted.org/packages/d7/de/35d839aa69db737a3f9f3a45000ca24721834d40118652a5775d5eca8ebb/fonttools-4.59.2-cp311-cp311-win_amd64.whl", hash = "sha256:9836394e2f4ce5f9c0a7690ee93bd90aa1adc6b054f1a57b562c5d242c903104", size = 2265846 },
+    { url = "https://files.pythonhosted.org/packages/ba/3d/1f45db2df51e7bfa55492e8f23f383d372200be3a0ded4bf56a92753dd1f/fonttools-4.59.2-cp312-cp312-macosx_10_13_universal2.whl", hash = "sha256:82906d002c349cad647a7634b004825a7335f8159d0d035ae89253b4abf6f3ea", size = 2769711 },
+    { url = "https://files.pythonhosted.org/packages/29/df/cd236ab32a8abfd11558f296e064424258db5edefd1279ffdbcfd4fd8b76/fonttools-4.59.2-cp312-cp312-macosx_10_13_x86_64.whl", hash = "sha256:a10c1bd7644dc58f8862d8ba0cf9fb7fef0af01ea184ba6ce3f50ab7dfe74d5a", size = 2340225 },
+    { url = "https://files.pythonhosted.org/packages/98/12/b6f9f964fe6d4b4dd4406bcbd3328821c3de1f909ffc3ffa558fe72af48c/fonttools-4.59.2-cp312-cp312-manylinux1_x86_64.manylinux2014_x86_64.manylinux_2_17_x86_64.manylinux_2_5_x86_64.whl", hash = "sha256:738f31f23e0339785fd67652a94bc69ea49e413dfdb14dcb8c8ff383d249464e", size = 4912766 },
+    { url = "https://files.pythonhosted.org/packages/73/78/82bde2f2d2c306ef3909b927363170b83df96171f74e0ccb47ad344563cd/fonttools-4.59.2-cp312-cp312-manylinux2014_aarch64.manylinux_2_17_aarch64.manylinux_2_28_aarch64.whl", hash = "sha256:0ec99f9bdfee9cdb4a9172f9e8fd578cce5feb231f598909e0aecf5418da4f25", size = 4955178 },
+    { url = "https://files.pythonhosted.org/packages/92/77/7de766afe2d31dda8ee46d7e479f35c7d48747e558961489a2d6e3a02bd4/fonttools-4.59.2-cp312-cp312-musllinux_1_2_aarch64.whl", hash = "sha256:0476ea74161322e08c7a982f83558a2b81b491509984523a1a540baf8611cc31", size = 4897898 },
+    { url = "https://files.pythonhosted.org/packages/c5/77/ce0e0b905d62a06415fda9f2b2e109a24a5db54a59502b769e9e297d2242/fonttools-4.59.2-cp312-cp312-musllinux_1_2_x86_64.whl", hash = "sha256:95922a922daa1f77cc72611747c156cfb38030ead72436a2c551d30ecef519b9", size = 5049144 },
+    { url = "https://files.pythonhosted.org/packages/d9/ea/870d93aefd23fff2e07cbeebdc332527868422a433c64062c09d4d5e7fe6/fonttools-4.59.2-cp312-cp312-win32.whl", hash = "sha256:39ad9612c6a622726a6a130e8ab15794558591f999673f1ee7d2f3d30f6a3e1c", size = 2206473 },
+    { url = "https://files.pythonhosted.org/packages/61/c4/e44bad000c4a4bb2e9ca11491d266e857df98ab6d7428441b173f0fe2517/fonttools-4.59.2-cp312-cp312-win_amd64.whl", hash = "sha256:980fd7388e461b19a881d35013fec32c713ffea1fc37aef2f77d11f332dfd7da", size = 2254706 },
+    { url = "https://files.pythonhosted.org/packages/65/a4/d2f7be3c86708912c02571db0b550121caab8cd88a3c0aacb9cfa15ea66e/fonttools-4.59.2-py3-none-any.whl", hash = "sha256:8bd0f759020e87bb5d323e6283914d9bf4ae35a7307dafb2cbd1e379e720ad37", size = 1132315 },
+]
+
 [[package]]
 name = "frozenlist"
 version = "1.7.0"
@@ -991,6 +1111,11 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/2f/e0/014d5d9d7a4564cf1c40b5039bc882db69fd881111e03ab3657ac0b218e2/fsspec-2025.7.0-py3-none-any.whl", hash = "sha256:8b012e39f63c7d5f10474de957f3ab793b47b45ae7d39f2fb735f8bbe25c0e21", size = 199597 },
 ]
 
+[package.optional-dependencies]
+http = [
+    { name = "aiohttp" },
+]
+
 [[package]]
 name = "google-crc32c"
 version = "1.7.1"
@@ -1255,6 +1380,19 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/f0/0f/310fb31e39e2d734ccaa2c0fb981ee41f7bd5056ce9bc29b2248bd569169/humanfriendly-10.0-py2.py3-none-any.whl", hash = "sha256:1697e1a8a8f550fd43c2865cd84542fc175a61dcb779b6fee18cf6b6ccba1477", size = 86794 },
 ]
 
+[[package]]
+name = "hyperpyyaml"
+version = "1.2.2"
+source = { registry = "https://pypi.org/simple" }
+dependencies = [
+    { name = "pyyaml" },
+    { name = "ruamel-yaml" },
+]
+sdist = { url = "https://files.pythonhosted.org/packages/52/e3/3ac46d9a662b037f699a6948b39c8d03bfcff0b592335d5953ba0c55d453/HyperPyYAML-1.2.2.tar.gz", hash = "sha256:bdb734210d18770a262f500fe5755c7a44a5d3b91521b06e24f7a00a36ee0f87", size = 17085 }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/33/c9/751b6401887f4b50f9307cc1e53d287b3dc77c375c126aeb6335aff73ccb/HyperPyYAML-1.2.2-py3-none-any.whl", hash = "sha256:3c5864bdc8864b2f0fbd7bc495e7e8fdf2dfd5dd80116f72da27ca96a128bdeb", size = 16118 },
+]
+
 [[package]]
 name = "icalendar"
 version = "6.3.1"
@@ -1397,6 +1535,55 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/01/0e/b27cdbaccf30b890c40ed1da9fd4a3593a5cf94dae54fb34f8a4b74fcd3f/jsonschema_specifications-2025.4.1-py3-none-any.whl", hash = "sha256:4653bffbd6584f7de83a67e0d620ef16900b390ddc7939d56684d6c81e33f1af", size = 18437 },
 ]
 
+[[package]]
+name = "julius"
+version = "0.2.7"
+source = { registry = "https://pypi.org/simple" }
+dependencies = [
+    { name = "torch", version = "2.8.0", source = { registry = "https://download.pytorch.org/whl/cpu" }, marker = "sys_platform == 'darwin'" },
+    { name = "torch", version = "2.8.0+cpu", source = { registry = "https://download.pytorch.org/whl/cpu" }, marker = "sys_platform != 'darwin'" },
+]
+sdist = { url = "https://files.pythonhosted.org/packages/a1/19/c9e1596b5572c786b93428d0904280e964c930fae7e6c9368ed9e1b63922/julius-0.2.7.tar.gz", hash = "sha256:3c0f5f5306d7d6016fcc95196b274cae6f07e2c9596eed314e4e7641554fbb08", size = 59640 }
+
+[[package]]
+name = "kiwisolver"
+version = "1.4.9"
+source = { registry = "https://pypi.org/simple" }
+sdist = { url = "https://files.pythonhosted.org/packages/5c/3c/85844f1b0feb11ee581ac23fe5fce65cd049a200c1446708cc1b7f922875/kiwisolver-1.4.9.tar.gz", hash = "sha256:c3b22c26c6fd6811b0ae8363b95ca8ce4ea3c202d3d0975b2914310ceb1bcc4d", size = 97564 }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/6f/ab/c80b0d5a9d8a1a65f4f815f2afff9798b12c3b9f31f1d304dd233dd920e2/kiwisolver-1.4.9-cp311-cp311-macosx_10_9_universal2.whl", hash = "sha256:eb14a5da6dc7642b0f3a18f13654847cd8b7a2550e2645a5bda677862b03ba16", size = 124167 },
+    { url = "https://files.pythonhosted.org/packages/a0/c0/27fe1a68a39cf62472a300e2879ffc13c0538546c359b86f149cc19f6ac3/kiwisolver-1.4.9-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:39a219e1c81ae3b103643d2aedb90f1ef22650deb266ff12a19e7773f3e5f089", size = 66579 },
+    { url = "https://files.pythonhosted.org/packages/31/a2/a12a503ac1fd4943c50f9822678e8015a790a13b5490354c68afb8489814/kiwisolver-1.4.9-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:2405a7d98604b87f3fc28b1716783534b1b4b8510d8142adca34ee0bc3c87543", size = 65309 },
+    { url = "https://files.pythonhosted.org/packages/66/e1/e533435c0be77c3f64040d68d7a657771194a63c279f55573188161e81ca/kiwisolver-1.4.9-cp311-cp311-manylinux2014_x86_64.manylinux_2_17_x86_64.whl", hash = "sha256:dc1ae486f9abcef254b5618dfb4113dd49f94c68e3e027d03cf0143f3f772b61", size = 1435596 },
+    { url = "https://files.pythonhosted.org/packages/67/1e/51b73c7347f9aabdc7215aa79e8b15299097dc2f8e67dee2b095faca9cb0/kiwisolver-1.4.9-cp311-cp311-manylinux_2_24_aarch64.manylinux_2_28_aarch64.whl", hash = "sha256:8a1f570ce4d62d718dce3f179ee78dac3b545ac16c0c04bb363b7607a949c0d1", size = 1246548 },
+    { url = "https://files.pythonhosted.org/packages/21/aa/72a1c5d1e430294f2d32adb9542719cfb441b5da368d09d268c7757af46c/kiwisolver-1.4.9-cp311-cp311-manylinux_2_24_ppc64le.manylinux_2_28_ppc64le.whl", hash = "sha256:cb27e7b78d716c591e88e0a09a2139c6577865d7f2e152488c2cc6257f460872", size = 1263618 },
+    { url = "https://files.pythonhosted.org/packages/a3/af/db1509a9e79dbf4c260ce0cfa3903ea8945f6240e9e59d1e4deb731b1a40/kiwisolver-1.4.9-cp311-cp311-manylinux_2_24_s390x.manylinux_2_28_s390x.whl", hash = "sha256:15163165efc2f627eb9687ea5f3a28137217d217ac4024893d753f46bce9de26", size = 1317437 },
+    { url = "https://files.pythonhosted.org/packages/e0/f2/3ea5ee5d52abacdd12013a94130436e19969fa183faa1e7c7fbc89e9a42f/kiwisolver-1.4.9-cp311-cp311-musllinux_1_2_aarch64.whl", hash = "sha256:bdee92c56a71d2b24c33a7d4c2856bd6419d017e08caa7802d2963870e315028", size = 2195742 },
+    { url = "https://files.pythonhosted.org/packages/6f/9b/1efdd3013c2d9a2566aa6a337e9923a00590c516add9a1e89a768a3eb2fc/kiwisolver-1.4.9-cp311-cp311-musllinux_1_2_ppc64le.whl", hash = "sha256:412f287c55a6f54b0650bd9b6dce5aceddb95864a1a90c87af16979d37c89771", size = 2290810 },
+    { url = "https://files.pythonhosted.org/packages/fb/e5/cfdc36109ae4e67361f9bc5b41323648cb24a01b9ade18784657e022e65f/kiwisolver-1.4.9-cp311-cp311-musllinux_1_2_s390x.whl", hash = "sha256:2c93f00dcba2eea70af2be5f11a830a742fe6b579a1d4e00f47760ef13be247a", size = 2461579 },
+    { url = "https://files.pythonhosted.org/packages/62/86/b589e5e86c7610842213994cdea5add00960076bef4ae290c5fa68589cac/kiwisolver-1.4.9-cp311-cp311-musllinux_1_2_x86_64.whl", hash = "sha256:f117e1a089d9411663a3207ba874f31be9ac8eaa5b533787024dc07aeb74f464", size = 2268071 },
+    { url = "https://files.pythonhosted.org/packages/3b/c6/f8df8509fd1eee6c622febe54384a96cfaf4d43bf2ccec7a0cc17e4715c9/kiwisolver-1.4.9-cp311-cp311-win_amd64.whl", hash = "sha256:be6a04e6c79819c9a8c2373317d19a96048e5a3f90bec587787e86a1153883c2", size = 73840 },
+    { url = "https://files.pythonhosted.org/packages/e2/2d/16e0581daafd147bc11ac53f032a2b45eabac897f42a338d0a13c1e5c436/kiwisolver-1.4.9-cp311-cp311-win_arm64.whl", hash = "sha256:0ae37737256ba2de764ddc12aed4956460277f00c4996d51a197e72f62f5eec7", size = 65159 },
+    { url = "https://files.pythonhosted.org/packages/86/c9/13573a747838aeb1c76e3267620daa054f4152444d1f3d1a2324b78255b5/kiwisolver-1.4.9-cp312-cp312-macosx_10_13_universal2.whl", hash = "sha256:ac5a486ac389dddcc5bef4f365b6ae3ffff2c433324fb38dd35e3fab7c957999", size = 123686 },
+    { url = "https://files.pythonhosted.org/packages/51/ea/2ecf727927f103ffd1739271ca19c424d0e65ea473fbaeea1c014aea93f6/kiwisolver-1.4.9-cp312-cp312-macosx_10_13_x86_64.whl", hash = "sha256:f2ba92255faa7309d06fe44c3a4a97efe1c8d640c2a79a5ef728b685762a6fd2", size = 66460 },
+    { url = "https://files.pythonhosted.org/packages/5b/5a/51f5464373ce2aeb5194508298a508b6f21d3867f499556263c64c621914/kiwisolver-1.4.9-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:4a2899935e724dd1074cb568ce7ac0dce28b2cd6ab539c8e001a8578eb106d14", size = 64952 },
+    { url = "https://files.pythonhosted.org/packages/70/90/6d240beb0f24b74371762873e9b7f499f1e02166a2d9c5801f4dbf8fa12e/kiwisolver-1.4.9-cp312-cp312-manylinux2014_x86_64.manylinux_2_17_x86_64.whl", hash = "sha256:f6008a4919fdbc0b0097089f67a1eb55d950ed7e90ce2cc3e640abadd2757a04", size = 1474756 },
+    { url = "https://files.pythonhosted.org/packages/12/42/f36816eaf465220f683fb711efdd1bbf7a7005a2473d0e4ed421389bd26c/kiwisolver-1.4.9-cp312-cp312-manylinux_2_24_aarch64.manylinux_2_28_aarch64.whl", hash = "sha256:67bb8b474b4181770f926f7b7d2f8c0248cbcb78b660fdd41a47054b28d2a752", size = 1276404 },
+    { url = "https://files.pythonhosted.org/packages/2e/64/bc2de94800adc830c476dce44e9b40fd0809cddeef1fde9fcf0f73da301f/kiwisolver-1.4.9-cp312-cp312-manylinux_2_24_ppc64le.manylinux_2_28_ppc64le.whl", hash = "sha256:2327a4a30d3ee07d2fbe2e7933e8a37c591663b96ce42a00bc67461a87d7df77", size = 1294410 },
+    { url = "https://files.pythonhosted.org/packages/5f/42/2dc82330a70aa8e55b6d395b11018045e58d0bb00834502bf11509f79091/kiwisolver-1.4.9-cp312-cp312-manylinux_2_24_s390x.manylinux_2_28_s390x.whl", hash = "sha256:7a08b491ec91b1d5053ac177afe5290adacf1f0f6307d771ccac5de30592d198", size = 1343631 },
+    { url = "https://files.pythonhosted.org/packages/22/fd/f4c67a6ed1aab149ec5a8a401c323cee7a1cbe364381bb6c9c0d564e0e20/kiwisolver-1.4.9-cp312-cp312-musllinux_1_2_aarch64.whl", hash = "sha256:d8fc5c867c22b828001b6a38d2eaeb88160bf5783c6cb4a5e440efc981ce286d", size = 2224963 },
+    { url = "https://files.pythonhosted.org/packages/45/aa/76720bd4cb3713314677d9ec94dcc21ced3f1baf4830adde5bb9b2430a5f/kiwisolver-1.4.9-cp312-cp312-musllinux_1_2_ppc64le.whl", hash = "sha256:3b3115b2581ea35bb6d1f24a4c90af37e5d9b49dcff267eeed14c3893c5b86ab", size = 2321295 },
+    { url = "https://files.pythonhosted.org/packages/80/19/d3ec0d9ab711242f56ae0dc2fc5d70e298bb4a1f9dfab44c027668c673a1/kiwisolver-1.4.9-cp312-cp312-musllinux_1_2_s390x.whl", hash = "sha256:858e4c22fb075920b96a291928cb7dea5644e94c0ee4fcd5af7e865655e4ccf2", size = 2487987 },
+    { url = "https://files.pythonhosted.org/packages/39/e9/61e4813b2c97e86b6fdbd4dd824bf72d28bcd8d4849b8084a357bc0dd64d/kiwisolver-1.4.9-cp312-cp312-musllinux_1_2_x86_64.whl", hash = "sha256:ed0fecd28cc62c54b262e3736f8bb2512d8dcfdc2bcf08be5f47f96bf405b145", size = 2291817 },
+    { url = "https://files.pythonhosted.org/packages/a0/41/85d82b0291db7504da3c2defe35c9a8a5c9803a730f297bd823d11d5fb77/kiwisolver-1.4.9-cp312-cp312-win_amd64.whl", hash = "sha256:f68208a520c3d86ea51acf688a3e3002615a7f0238002cccc17affecc86a8a54", size = 73895 },
+    { url = "https://files.pythonhosted.org/packages/e2/92/5f3068cf15ee5cb624a0c7596e67e2a0bb2adee33f71c379054a491d07da/kiwisolver-1.4.9-cp312-cp312-win_arm64.whl", hash = "sha256:2c1a4f57df73965f3f14df20b80ee29e6a7930a57d2d9e8491a25f676e197c60", size = 64992 },
+    { url = "https://files.pythonhosted.org/packages/a3/0f/36d89194b5a32c054ce93e586d4049b6c2c22887b0eb229c61c68afd3078/kiwisolver-1.4.9-pp311-pypy311_pp73-macosx_10_15_x86_64.whl", hash = "sha256:720e05574713db64c356e86732c0f3c5252818d05f9df320f0ad8380641acea5", size = 60104 },
+    { url = "https://files.pythonhosted.org/packages/52/ba/4ed75f59e4658fd21fe7dde1fee0ac397c678ec3befba3fe6482d987af87/kiwisolver-1.4.9-pp311-pypy311_pp73-macosx_11_0_arm64.whl", hash = "sha256:17680d737d5335b552994a2008fab4c851bcd7de33094a82067ef3a576ff02fa", size = 58592 },
+    { url = "https://files.pythonhosted.org/packages/33/01/a8ea7c5ea32a9b45ceeaee051a04c8ed4320f5add3c51bfa20879b765b70/kiwisolver-1.4.9-pp311-pypy311_pp73-manylinux2014_x86_64.manylinux_2_17_x86_64.whl", hash = "sha256:85b5352f94e490c028926ea567fc569c52ec79ce131dadb968d3853e809518c2", size = 80281 },
+    { url = "https://files.pythonhosted.org/packages/da/e3/dbd2ecdce306f1d07a1aaf324817ee993aab7aee9db47ceac757deabafbe/kiwisolver-1.4.9-pp311-pypy311_pp73-manylinux_2_24_aarch64.manylinux_2_28_aarch64.whl", hash = "sha256:464415881e4801295659462c49461a24fb107c140de781d55518c4b80cb6790f", size = 78009 },
+    { url = "https://files.pythonhosted.org/packages/da/e9/0d4add7873a73e462aeb45c036a2dead2562b825aa46ba326727b3f31016/kiwisolver-1.4.9-pp311-pypy311_pp73-win_amd64.whl", hash = "sha256:fb940820c63a9590d31d88b815e7a3aa5915cad3ce735ab45f0c730b39547de1", size = 73929 },
+]
+
 [[package]]
 name = "kombu"
 version = "5.5.4"
@@ -1459,6 +1646,41 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/dc/1e/408fd10217eac0e43aea0604be22b4851a09e03d761d44d4ea12089dd70e/levenshtein-0.27.1-pp311-pypy311_pp73-win_amd64.whl", hash = "sha256:7987ef006a3cf56a4532bd4c90c2d3b7b4ca9ad3bf8ae1ee5713c4a3bdfda913", size = 98045 },
 ]
 
+[[package]]
+name = "lightning"
+version = "2.5.5"
+source = { registry = "https://pypi.org/simple" }
+dependencies = [
+    { name = "fsspec", extra = ["http"] },
+    { name = "lightning-utilities" },
+    { name = "packaging" },
+    { name = "pytorch-lightning" },
+    { name = "pyyaml" },
+    { name = "torch", version = "2.8.0", source = { registry = "https://download.pytorch.org/whl/cpu" }, marker = "sys_platform == 'darwin'" },
+    { name = "torch", version = "2.8.0+cpu", source = { registry = "https://download.pytorch.org/whl/cpu" }, marker = "sys_platform != 'darwin'" },
+    { name = "torchmetrics" },
+    { name = "tqdm" },
+    { name = "typing-extensions" },
+]
+sdist = { url = "https://files.pythonhosted.org/packages/0f/dd/86bb3bebadcdbc6e6e5a63657f0a03f74cd065b5ea965896679f76fec0b4/lightning-2.5.5.tar.gz", hash = "sha256:4d3d66c5b1481364a7e6a1ce8ddde1777a04fa740a3145ec218a9941aed7dd30", size = 640770 }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/2e/d0/4b4fbafc3b18df91207a6e46782d9fd1905f9f45cb2c3b8dfbb239aef781/lightning-2.5.5-py3-none-any.whl", hash = "sha256:69eb248beadd7b600bf48eff00a0ec8af171ec7a678d23787c4aedf12e225e8f", size = 828490 },
+]
+
+[[package]]
+name = "lightning-utilities"
+version = "0.15.2"
+source = { registry = "https://pypi.org/simple" }
+dependencies = [
+    { name = "packaging" },
+    { name = "setuptools" },
+    { name = "typing-extensions" },
+]
+sdist = { url = "https://files.pythonhosted.org/packages/b8/39/6fc58ca81492db047149b4b8fd385aa1bfb8c28cd7cacb0c7eb0c44d842f/lightning_utilities-0.15.2.tar.gz", hash = "sha256:cdf12f530214a63dacefd713f180d1ecf5d165338101617b4742e8f22c032e24", size = 31090 }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/de/73/3d757cb3fc16f0f9794dd289bcd0c4a031d9cf54d8137d6b984b2d02edf3/lightning_utilities-0.15.2-py3-none-any.whl", hash = "sha256:ad3ab1703775044bbf880dbf7ddaaac899396c96315f3aa1779cec9d618a9841", size = 29431 },
+]
+
 [[package]]
 name = "llama-cloud"
 version = "0.1.32"
@@ -1806,6 +2028,42 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/34/75/51952c7b2d3873b44a0028b1bd26a25078c18f92f256608e8d1dc61b39fd/marshmallow-3.26.1-py3-none-any.whl", hash = "sha256:3350409f20a70a7e4e11a27661187b77cdcaeb20abca41c1454fe33636bea09c", size = 50878 },
 ]
 
+[[package]]
+name = "matplotlib"
+version = "3.10.6"
+source = { registry = "https://pypi.org/simple" }
+dependencies = [
+    { name = "contourpy" },
+    { name = "cycler" },
+    { name = "fonttools" },
+    { name = "kiwisolver" },
+    { name = "numpy" },
+    { name = "packaging" },
+    { name = "pillow" },
+    { name = "pyparsing" },
+    { name = "python-dateutil" },
+]
+sdist = { url = "https://files.pythonhosted.org/packages/a0/59/c3e6453a9676ffba145309a73c462bb407f4400de7de3f2b41af70720a3c/matplotlib-3.10.6.tar.gz", hash = "sha256:ec01b645840dd1996df21ee37f208cd8ba57644779fa20464010638013d3203c", size = 34804264 }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/80/d6/5d3665aa44c49005aaacaa68ddea6fcb27345961cd538a98bb0177934ede/matplotlib-3.10.6-cp311-cp311-macosx_10_12_x86_64.whl", hash = "sha256:905b60d1cb0ee604ce65b297b61cf8be9f4e6cfecf95a3fe1c388b5266bc8f4f", size = 8257527 },
+    { url = "https://files.pythonhosted.org/packages/8c/af/30ddefe19ca67eebd70047dabf50f899eaff6f3c5e6a1a7edaecaf63f794/matplotlib-3.10.6-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:7bac38d816637343e53d7185d0c66677ff30ffb131044a81898b5792c956ba76", size = 8119583 },
+    { url = "https://files.pythonhosted.org/packages/d3/29/4a8650a3dcae97fa4f375d46efcb25920d67b512186f8a6788b896062a81/matplotlib-3.10.6-cp311-cp311-manylinux2014_x86_64.manylinux_2_17_x86_64.whl", hash = "sha256:942a8de2b5bfff1de31d95722f702e2966b8a7e31f4e68f7cd963c7cd8861cf6", size = 8692682 },
+    { url = "https://files.pythonhosted.org/packages/aa/d3/b793b9cb061cfd5d42ff0f69d1822f8d5dbc94e004618e48a97a8373179a/matplotlib-3.10.6-cp311-cp311-manylinux_2_27_aarch64.manylinux_2_28_aarch64.whl", hash = "sha256:a3276c85370bc0dfca051ec65c5817d1e0f8f5ce1b7787528ec8ed2d524bbc2f", size = 9521065 },
+    { url = "https://files.pythonhosted.org/packages/f7/c5/53de5629f223c1c66668d46ac2621961970d21916a4bc3862b174eb2a88f/matplotlib-3.10.6-cp311-cp311-musllinux_1_2_x86_64.whl", hash = "sha256:9df5851b219225731f564e4b9e7f2ac1e13c9e6481f941b5631a0f8e2d9387ce", size = 9576888 },
+    { url = "https://files.pythonhosted.org/packages/fc/8e/0a18d6d7d2d0a2e66585032a760d13662e5250c784d53ad50434e9560991/matplotlib-3.10.6-cp311-cp311-win_amd64.whl", hash = "sha256:abb5d9478625dd9c9eb51a06d39aae71eda749ae9b3138afb23eb38824026c7e", size = 8115158 },
+    { url = "https://files.pythonhosted.org/packages/07/b3/1a5107bb66c261e23b9338070702597a2d374e5aa7004b7adfc754fbed02/matplotlib-3.10.6-cp311-cp311-win_arm64.whl", hash = "sha256:886f989ccfae63659183173bb3fced7fd65e9eb793c3cc21c273add368536951", size = 7992444 },
+    { url = "https://files.pythonhosted.org/packages/ea/1a/7042f7430055d567cc3257ac409fcf608599ab27459457f13772c2d9778b/matplotlib-3.10.6-cp312-cp312-macosx_10_13_x86_64.whl", hash = "sha256:31ca662df6a80bd426f871105fdd69db7543e28e73a9f2afe80de7e531eb2347", size = 8272404 },
+    { url = "https://files.pythonhosted.org/packages/a9/5d/1d5f33f5b43f4f9e69e6a5fe1fb9090936ae7bc8e2ff6158e7a76542633b/matplotlib-3.10.6-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:1678bb61d897bb4ac4757b5ecfb02bfb3fddf7f808000fb81e09c510712fda75", size = 8128262 },
+    { url = "https://files.pythonhosted.org/packages/67/c3/135fdbbbf84e0979712df58e5e22b4f257b3f5e52a3c4aacf1b8abec0d09/matplotlib-3.10.6-cp312-cp312-manylinux2014_x86_64.manylinux_2_17_x86_64.whl", hash = "sha256:56cd2d20842f58c03d2d6e6c1f1cf5548ad6f66b91e1e48f814e4fb5abd1cb95", size = 8697008 },
+    { url = "https://files.pythonhosted.org/packages/9c/be/c443ea428fb2488a3ea7608714b1bd85a82738c45da21b447dc49e2f8e5d/matplotlib-3.10.6-cp312-cp312-manylinux_2_27_aarch64.manylinux_2_28_aarch64.whl", hash = "sha256:662df55604a2f9a45435566d6e2660e41efe83cd94f4288dfbf1e6d1eae4b0bb", size = 9530166 },
+    { url = "https://files.pythonhosted.org/packages/a9/35/48441422b044d74034aea2a3e0d1a49023f12150ebc58f16600132b9bbaf/matplotlib-3.10.6-cp312-cp312-musllinux_1_2_x86_64.whl", hash = "sha256:08f141d55148cd1fc870c3387d70ca4df16dee10e909b3b038782bd4bda6ea07", size = 9593105 },
+    { url = "https://files.pythonhosted.org/packages/45/c3/994ef20eb4154ab84cc08d033834555319e4af970165e6c8894050af0b3c/matplotlib-3.10.6-cp312-cp312-win_amd64.whl", hash = "sha256:590f5925c2d650b5c9d813c5b3b5fc53f2929c3f8ef463e4ecfa7e052044fb2b", size = 8122784 },
+    { url = "https://files.pythonhosted.org/packages/57/b8/5c85d9ae0e40f04e71bedb053aada5d6bab1f9b5399a0937afb5d6b02d98/matplotlib-3.10.6-cp312-cp312-win_arm64.whl", hash = "sha256:f44c8d264a71609c79a78d50349e724f5d5fc3684ead7c2a473665ee63d868aa", size = 7992823 },
+    { url = "https://files.pythonhosted.org/packages/12/bb/02c35a51484aae5f49bd29f091286e7af5f3f677a9736c58a92b3c78baeb/matplotlib-3.10.6-pp311-pypy311_pp73-macosx_10_15_x86_64.whl", hash = "sha256:f2d684c3204fa62421bbf770ddfebc6b50130f9cad65531eeba19236d73bb488", size = 8252296 },
+    { url = "https://files.pythonhosted.org/packages/7d/85/41701e3092005aee9a2445f5ee3904d9dbd4a7df7a45905ffef29b7ef098/matplotlib-3.10.6-pp311-pypy311_pp73-macosx_11_0_arm64.whl", hash = "sha256:6f4a69196e663a41d12a728fab8751177215357906436804217d6d9cf0d4d6cf", size = 8116749 },
+    { url = "https://files.pythonhosted.org/packages/16/53/8d8fa0ea32a8c8239e04d022f6c059ee5e1b77517769feccd50f1df43d6d/matplotlib-3.10.6-pp311-pypy311_pp73-manylinux2014_x86_64.manylinux_2_17_x86_64.whl", hash = "sha256:4d6ca6ef03dfd269f4ead566ec6f3fb9becf8dab146fb999022ed85ee9f6b3eb", size = 8693933 },
+]
+
 [[package]]
 name = "mdurl"
 version = "0.1.2"
@@ -1947,6 +2205,19 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/48/6b/1c6b515a83d5564b1698a61efa245727c8feecf308f4091f565988519d20/numpy-2.3.1-pp311-pypy311_pp73-win_amd64.whl", hash = "sha256:e610832418a2bc09d974cc9fecebfa51e9532d6190223bc5ef6a7402ebf3b5cb", size = 12927246 },
 ]
 
+[[package]]
+name = "omegaconf"
+version = "2.3.0"
+source = { registry = "https://pypi.org/simple" }
+dependencies = [
+    { name = "antlr4-python3-runtime" },
+    { name = "pyyaml" },
+]
+sdist = { url = "https://files.pythonhosted.org/packages/09/48/6388f1bb9da707110532cb70ec4d2822858ddfb44f1cdf1233c20a80ea4b/omegaconf-2.3.0.tar.gz", hash = "sha256:d5d4b6d29955cc50ad50c46dc269bcd92c6e00f5f90d23ab5fee7bfca4ba4cc7", size = 3298120 }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/e3/94/1843518e420fa3ed6919835845df698c7e27e183cb997394e4a670973a65/omegaconf-2.3.0-py3-none-any.whl", hash = "sha256:7b4df175cdb08ba400f45cae3bdcae7ba8365db4d165fc65fd04b050ab63b46b", size = 79500 },
+]
+
 [[package]]
 name = "onnxruntime"
 version = "1.22.1"
@@ -1989,6 +2260,24 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/8a/91/1f1cf577f745e956b276a8b1d3d76fa7a6ee0c2b05db3b001b900f2c71db/openai-1.97.0-py3-none-any.whl", hash = "sha256:a1c24d96f4609f3f7f51c9e1c2606d97cc6e334833438659cfd687e9c972c610", size = 764953 },
 ]
 
+[[package]]
+name = "optuna"
+version = "4.5.0"
+source = { registry = "https://pypi.org/simple" }
+dependencies = [
+    { name = "alembic" },
+    { name = "colorlog" },
+    { name = "numpy" },
+    { name = "packaging" },
+    { name = "pyyaml" },
+    { name = "sqlalchemy" },
+    { name = "tqdm" },
+]
+sdist = { url = "https://files.pythonhosted.org/packages/53/a3/bcd1e5500de6ec794c085a277e5b624e60b4fac1790681d7cdbde25b93a2/optuna-4.5.0.tar.gz", hash = "sha256:264844da16dad744dea295057d8bc218646129c47567d52c35a201d9f99942ba", size = 472338 }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/7f/12/cba81286cbaf0f0c3f0473846cfd992cb240bdcea816bf2ef7de8ed0f744/optuna-4.5.0-py3-none-any.whl", hash = "sha256:5b8a783e84e448b0742501bc27195344a28d2c77bd2feef5b558544d954851b0", size = 400872 },
+]
+
 [[package]]
 name = "packaging"
 version = "25.0"
@@ -2090,6 +2379,15 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/54/20/4d324d65cc6d9205fabedc306948156824eb9f0ee1633355a8f7ec5c66bf/pluggy-1.6.0-py3-none-any.whl", hash = "sha256:e920276dd6813095e9377c0bc5566d94c932c33b27a3e3945d8389c374dd4746", size = 20538 },
 ]
 
+[[package]]
+name = "primepy"
+version = "1.3"
+source = { registry = "https://pypi.org/simple" }
+sdist = { url = "https://files.pythonhosted.org/packages/35/77/0cfa1b4697cfb5336f3a96e8bc73327f64610be3a64c97275f1801afb395/primePy-1.3.tar.gz", hash = "sha256:25fd7e25344b0789a5984c75d89f054fcf1f180bef20c998e4befbac92de4669", size = 3914 }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/74/c1/bb7e334135859c3a92ec399bc89293ea73f28e815e35b43929c8db6af030/primePy-1.3-py3-none-any.whl", hash = "sha256:5ed443718765be9bf7e2ff4c56cdff71b42140a15b39d054f9d99f0009e2317a", size = 4040 },
+]
+
 [[package]]
 name = "prometheus-client"
 version = "0.22.1"
@@ -2226,6 +2524,109 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/92/29/06261ea000e2dc1e22907dbbc483a1093665509ea586b29b8986a0e56733/psycopg2_binary-2.9.10-cp312-cp312-win_amd64.whl", hash = "sha256:18c5ee682b9c6dd3696dad6e54cc7ff3a1a9020df6a5c0f861ef8bfd338c3ca0", size = 1164031 },
 ]
 
+[[package]]
+name = "pyannote-audio"
+version = "3.3.2"
+source = { registry = "https://pypi.org/simple" }
+dependencies = [
+    { name = "asteroid-filterbanks" },
+    { name = "einops" },
+    { name = "huggingface-hub" },
+    { name = "lightning" },
+    { name = "omegaconf" },
+    { name = "pyannote-core" },
+    { name = "pyannote-database" },
+    { name = "pyannote-metrics" },
+    { name = "pyannote-pipeline" },
+    { name = "pytorch-metric-learning" },
+    { name = "rich" },
+    { name = "semver" },
+    { name = "soundfile" },
+    { name = "speechbrain" },
+    { name = "tensorboardx" },
+    { name = "torch", version = "2.8.0", source = { registry = "https://download.pytorch.org/whl/cpu" }, marker = "sys_platform == 'darwin'" },
+    { name = "torch", version = "2.8.0+cpu", source = { registry = "https://download.pytorch.org/whl/cpu" }, marker = "sys_platform != 'darwin'" },
+    { name = "torch-audiomentations" },
+    { name = "torchaudio", version = "2.8.0", source = { registry = "https://download.pytorch.org/whl/cpu" }, marker = "(platform_machine == 'aarch64' and sys_platform == 'linux') or sys_platform == 'darwin'" },
+    { name = "torchaudio", version = "2.8.0+cpu", source = { registry = "https://download.pytorch.org/whl/cpu" }, marker = "(platform_machine != 'aarch64' and sys_platform == 'linux') or (sys_platform != 'darwin' and sys_platform != 'linux')" },
+    { name = "torchmetrics" },
+]
+sdist = { url = "https://files.pythonhosted.org/packages/e9/00/3b96ca7ad0641e4f64cfaa2af153dc7da0998ff972280e1c1681b1fcc243/pyannote_audio-3.3.2.tar.gz", hash = "sha256:b2115e86b0db5faedb9f36ee1a150cebd07f7758e65e815accdac1a12ca9c777", size = 13664309 }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/17/e6/76049470d90217f9a15a34abf3e92d782cabc3fb4ab27515c9baaa5495d1/pyannote.audio-3.3.2-py2.py3-none-any.whl", hash = "sha256:599c694acd5d193215147ff82d0bf638bb191204ed502bd9fde8ff582e20aa1c", size = 898707 },
+    { url = "https://files.pythonhosted.org/packages/b7/9a/98a8992727e762b031ed30451d5726ece46cf8bb7b872a9dba5cef011e5d/pyannote_audio-3.3.2-py2.py3-none-any.whl", hash = "sha256:23e0dcedda920cb2e154e146bcd9663289ee7942d0e012663dad76f2e571ebeb", size = 897827 },
+]
+
+[[package]]
+name = "pyannote-core"
+version = "5.0.0"
+source = { registry = "https://pypi.org/simple" }
+dependencies = [
+    { name = "numpy" },
+    { name = "scipy" },
+    { name = "sortedcontainers" },
+    { name = "typing-extensions" },
+]
+sdist = { url = "https://files.pythonhosted.org/packages/65/03/feaf7534206f02c75baf151ce4b8c322b402a6f477c2be82f69d9269cbe6/pyannote.core-5.0.0.tar.gz", hash = "sha256:1a55bcc8bd680ba6be5fa53efa3b6f3d2cdd67144c07b6b4d8d66d5cb0d2096f", size = 59247 }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/84/c4/370bc8ba66815a5832ece753a1009388bb07ea353d21c83f2d5a1a436f2c/pyannote.core-5.0.0-py3-none-any.whl", hash = "sha256:04920a6754492242ce0dc6017545595ab643870fe69a994f20c1a5f2da0544d0", size = 58475 },
+]
+
+[[package]]
+name = "pyannote-database"
+version = "5.1.3"
+source = { registry = "https://pypi.org/simple" }
+dependencies = [
+    { name = "pandas" },
+    { name = "pyannote-core" },
+    { name = "pyyaml" },
+    { name = "typer" },
+]
+sdist = { url = "https://files.pythonhosted.org/packages/a9/ae/de36413d69a46be87cb612ebbcdc4eacbeebce3bc809124603e44a88fe26/pyannote.database-5.1.3.tar.gz", hash = "sha256:0eaf64c1cc506718de60d2d702f1359b1ae7ff252ee3e4799f1c5e378cd52c31", size = 49957 }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/a1/64/92d51a3a05615ba58be8ba62a43f9f9f952d9f3646f7e4fb7826e5a3a24e/pyannote.database-5.1.3-py3-none-any.whl", hash = "sha256:37887844c7dfbcc075cb591eddc00aff45fae1ed905344e1f43e0090e63bd40a", size = 48127 },
+]
+
+[[package]]
+name = "pyannote-metrics"
+version = "3.2.1"
+source = { registry = "https://pypi.org/simple" }
+dependencies = [
+    { name = "docopt" },
+    { name = "matplotlib" },
+    { name = "numpy" },
+    { name = "pandas" },
+    { name = "pyannote-core" },
+    { name = "pyannote-database" },
+    { name = "scikit-learn" },
+    { name = "scipy" },
+    { name = "sympy" },
+    { name = "tabulate" },
+]
+sdist = { url = "https://files.pythonhosted.org/packages/39/2b/6c5f01d3c49aa1c160765946e23782ca6436ae8b9bc514b56319ff5f16e7/pyannote.metrics-3.2.1.tar.gz", hash = "sha256:08024255a3550e96a8e9da4f5f4af326886548480de891414567c8900920ee5c", size = 49086 }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/6c/7d/035b370ab834b30e849fe9cd092b7bd7f321fcc4a2c56b84e96476b7ede5/pyannote.metrics-3.2.1-py3-none-any.whl", hash = "sha256:46be797cdade26c82773e5018659ae610145260069c7c5bf3d3c8a029ade8e22", size = 51386 },
+]
+
+[[package]]
+name = "pyannote-pipeline"
+version = "3.0.1"
+source = { registry = "https://pypi.org/simple" }
+dependencies = [
+    { name = "docopt" },
+    { name = "filelock" },
+    { name = "optuna" },
+    { name = "pyannote-core" },
+    { name = "pyannote-database" },
+    { name = "pyyaml" },
+    { name = "scikit-learn" },
+    { name = "tqdm" },
+]
+sdist = { url = "https://files.pythonhosted.org/packages/35/04/4bcfe0dd588577a188328b806f3a7213d8cead0ce5fe5784d01fd57df93f/pyannote.pipeline-3.0.1.tar.gz", hash = "sha256:021794e26a2cf5d8fb5bb1835951e71f5fac33eb14e23dfb7468e16b1b805151", size = 34486 }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/83/42/1bf7cbf061ed05c580bfb63bffdd3f3474cbd5c02bee4fac518eea9e9d9e/pyannote.pipeline-3.0.1-py3-none-any.whl", hash = "sha256:819bde4c4dd514f740f2373dfec794832b9fc8e346a35e43a7681625ee187393", size = 31517 },
+]
+
 [[package]]
 name = "pyasn1"
 version = "0.6.1"
@@ -2405,6 +2806,15 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/80/28/2659c02301b9500751f8d42f9a6632e1508aa5120de5e43042b8b30f8d5d/pyopenssl-25.1.0-py3-none-any.whl", hash = "sha256:2b11f239acc47ac2e5aca04fd7fa829800aeee22a2eb30d744572a157bd8a1ab", size = 56771 },
 ]
 
+[[package]]
+name = "pyparsing"
+version = "3.2.3"
+source = { registry = "https://pypi.org/simple" }
+sdist = { url = "https://files.pythonhosted.org/packages/bb/22/f1129e69d94ffff626bdb5c835506b3a5b4f3d070f17ea295e12c2c6f60f/pyparsing-3.2.3.tar.gz", hash = "sha256:b9c13f1ab8b3b542f72e28f634bad4de758ab3ce4546e4301970ad6fa77c38be", size = 1088608 }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/05/e7/df2285f3d08fee213f2d041540fa4fc9ca6c2d44cf36d3a035bf2a8d2bcc/pyparsing-3.2.3-py3-none-any.whl", hash = "sha256:a749938e02d6fd0b59b356ca504a24982314bb090c383e3cf201c95ef7e2bfcf", size = 111120 },
+]
+
 [[package]]
 name = "pypdf"
 version = "5.8.0"
@@ -2612,6 +3022,42 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/45/58/38b5afbc1a800eeea951b9285d3912613f2603bdf897a4ab0f4bd7f405fc/python_multipart-0.0.20-py3-none-any.whl", hash = "sha256:8a62d3a8335e06589fe01f2a3e178cdcc632f3fbe0d492ad9ee0ec35aab1f104", size = 24546 },
 ]
 
+[[package]]
+name = "pytorch-lightning"
+version = "2.5.5"
+source = { registry = "https://pypi.org/simple" }
+dependencies = [
+    { name = "fsspec", extra = ["http"] },
+    { name = "lightning-utilities" },
+    { name = "packaging" },
+    { name = "pyyaml" },
+    { name = "torch", version = "2.8.0", source = { registry = "https://download.pytorch.org/whl/cpu" }, marker = "sys_platform == 'darwin'" },
+    { name = "torch", version = "2.8.0+cpu", source = { registry = "https://download.pytorch.org/whl/cpu" }, marker = "sys_platform != 'darwin'" },
+    { name = "torchmetrics" },
+    { name = "tqdm" },
+    { name = "typing-extensions" },
+]
+sdist = { url = "https://files.pythonhosted.org/packages/16/78/bce84aab9a5b3b2e9d087d4f1a6be9b481adbfaac4903bc9daaaf09d49a3/pytorch_lightning-2.5.5.tar.gz", hash = "sha256:d6fc8173d1d6e49abfd16855ea05d2eb2415e68593f33d43e59028ecb4e64087", size = 643703 }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/04/f6/99a5c66478f469598dee25b0e29b302b5bddd4e03ed0da79608ac964056e/pytorch_lightning-2.5.5-py3-none-any.whl", hash = "sha256:0b533991df2353c0c6ea9ca10a7d0728b73631fd61f5a15511b19bee2aef8af0", size = 832431 },
+]
+
+[[package]]
+name = "pytorch-metric-learning"
+version = "2.9.0"
+source = { registry = "https://pypi.org/simple" }
+dependencies = [
+    { name = "numpy" },
+    { name = "scikit-learn" },
+    { name = "torch", version = "2.8.0", source = { registry = "https://download.pytorch.org/whl/cpu" }, marker = "sys_platform == 'darwin'" },
+    { name = "torch", version = "2.8.0+cpu", source = { registry = "https://download.pytorch.org/whl/cpu" }, marker = "sys_platform != 'darwin'" },
+    { name = "tqdm" },
+]
+sdist = { url = "https://files.pythonhosted.org/packages/9b/80/6e61b1a91debf4c1b47d441f9a9d7fe2aabcdd9575ed70b2811474eb95c3/pytorch-metric-learning-2.9.0.tar.gz", hash = "sha256:27a626caf5e2876a0fd666605a78cb67ef7597e25d7a68c18053dd503830701f", size = 84530 }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/46/7d/73ef5052f57b7720cad00e16598db3592a5ef4826745ffca67a2f085d4dc/pytorch_metric_learning-2.9.0-py3-none-any.whl", hash = "sha256:d51646006dc87168f00cf954785db133a4c5aac81253877248737aa42ef6432a", size = 127801 },
+]
+
 [[package]]
 name = "pytz"
 version = "2025.2"
@@ -2788,6 +3234,7 @@ evaluation = [
 ]
 local = [
     { name = "faster-whisper" },
+    { name = "pyannote-audio" },
 ]
 silero-vad = [
     { name = "silero-vad" },
@@ -2860,7 +3307,10 @@ evaluation = [
     { name = "pydantic", specifier = ">=2.1.1" },
     { name = "tqdm", specifier = ">=4.66.0" },
 ]
-local = [{ name = "faster-whisper", specifier = ">=0.10.0" }]
+local = [
+    { name = "faster-whisper", specifier = ">=0.10.0" },
+    { name = "pyannote-audio", specifier = ">=3.3.2" },
+]
 silero-vad = [
     { name = "silero-vad", specifier = ">=5.1.2" },
     { name = "torch", specifier = ">=2.8.0", index = "https://download.pytorch.org/whl/cpu" },
@@ -3064,6 +3514,44 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/64/8d/0133e4eb4beed9e425d9a98ed6e081a55d195481b7632472be1af08d2f6b/rsa-4.9.1-py3-none-any.whl", hash = "sha256:68635866661c6836b8d39430f97a996acbd61bfa49406748ea243539fe239762", size = 34696 },
 ]
 
+[[package]]
+name = "ruamel-yaml"
+version = "0.18.15"
+source = { registry = "https://pypi.org/simple" }
+dependencies = [
+    { name = "ruamel-yaml-clib", marker = "platform_python_implementation == 'CPython'" },
+]
+sdist = { url = "https://files.pythonhosted.org/packages/3e/db/f3950f5e5031b618aae9f423a39bf81a55c148aecd15a34527898e752cf4/ruamel.yaml-0.18.15.tar.gz", hash = "sha256:dbfca74b018c4c3fba0b9cc9ee33e53c371194a9000e694995e620490fd40700", size = 146865 }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/d1/e5/f2a0621f1781b76a38194acae72f01e37b1941470407345b6e8653ad7640/ruamel.yaml-0.18.15-py3-none-any.whl", hash = "sha256:148f6488d698b7a5eded5ea793a025308b25eca97208181b6a026037f391f701", size = 119702 },
+]
+
+[[package]]
+name = "ruamel-yaml-clib"
+version = "0.2.12"
+source = { registry = "https://pypi.org/simple" }
+sdist = { url = "https://files.pythonhosted.org/packages/20/84/80203abff8ea4993a87d823a5f632e4d92831ef75d404c9fc78d0176d2b5/ruamel.yaml.clib-0.2.12.tar.gz", hash = "sha256:6c8fbb13ec503f99a91901ab46e0b07ae7941cd527393187039aec586fdfd36f", size = 225315 }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/fb/8f/683c6ad562f558cbc4f7c029abcd9599148c51c54b5ef0f24f2638da9fbb/ruamel.yaml.clib-0.2.12-cp311-cp311-macosx_13_0_arm64.whl", hash = "sha256:4a6679521a58256a90b0d89e03992c15144c5f3858f40d7c18886023d7943db6", size = 132224 },
+    { url = "https://files.pythonhosted.org/packages/3c/d2/b79b7d695e2f21da020bd44c782490578f300dd44f0a4c57a92575758a76/ruamel.yaml.clib-0.2.12-cp311-cp311-manylinux2014_aarch64.whl", hash = "sha256:d84318609196d6bd6da0edfa25cedfbabd8dbde5140a0a23af29ad4b8f91fb1e", size = 641480 },
+    { url = "https://files.pythonhosted.org/packages/68/6e/264c50ce2a31473a9fdbf4fa66ca9b2b17c7455b31ef585462343818bd6c/ruamel.yaml.clib-0.2.12-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:bb43a269eb827806502c7c8efb7ae7e9e9d0573257a46e8e952f4d4caba4f31e", size = 739068 },
+    { url = "https://files.pythonhosted.org/packages/86/29/88c2567bc893c84d88b4c48027367c3562ae69121d568e8a3f3a8d363f4d/ruamel.yaml.clib-0.2.12-cp311-cp311-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:811ea1594b8a0fb466172c384267a4e5e367298af6b228931f273b111f17ef52", size = 703012 },
+    { url = "https://files.pythonhosted.org/packages/11/46/879763c619b5470820f0cd6ca97d134771e502776bc2b844d2adb6e37753/ruamel.yaml.clib-0.2.12-cp311-cp311-musllinux_1_1_i686.whl", hash = "sha256:cf12567a7b565cbf65d438dec6cfbe2917d3c1bdddfce84a9930b7d35ea59642", size = 704352 },
+    { url = "https://files.pythonhosted.org/packages/02/80/ece7e6034256a4186bbe50dee28cd032d816974941a6abf6a9d65e4228a7/ruamel.yaml.clib-0.2.12-cp311-cp311-musllinux_1_1_x86_64.whl", hash = "sha256:7dd5adc8b930b12c8fc5b99e2d535a09889941aa0d0bd06f4749e9a9397c71d2", size = 737344 },
+    { url = "https://files.pythonhosted.org/packages/f0/ca/e4106ac7e80efbabdf4bf91d3d32fc424e41418458251712f5672eada9ce/ruamel.yaml.clib-0.2.12-cp311-cp311-musllinux_1_2_aarch64.whl", hash = "sha256:1492a6051dab8d912fc2adeef0e8c72216b24d57bd896ea607cb90bb0c4981d3", size = 714498 },
+    { url = "https://files.pythonhosted.org/packages/67/58/b1f60a1d591b771298ffa0428237afb092c7f29ae23bad93420b1eb10703/ruamel.yaml.clib-0.2.12-cp311-cp311-win32.whl", hash = "sha256:bd0a08f0bab19093c54e18a14a10b4322e1eacc5217056f3c063bd2f59853ce4", size = 100205 },
+    { url = "https://files.pythonhosted.org/packages/b4/4f/b52f634c9548a9291a70dfce26ca7ebce388235c93588a1068028ea23fcc/ruamel.yaml.clib-0.2.12-cp311-cp311-win_amd64.whl", hash = "sha256:a274fb2cb086c7a3dea4322ec27f4cb5cc4b6298adb583ab0e211a4682f241eb", size = 118185 },
+    { url = "https://files.pythonhosted.org/packages/48/41/e7a405afbdc26af961678474a55373e1b323605a4f5e2ddd4a80ea80f628/ruamel.yaml.clib-0.2.12-cp312-cp312-macosx_14_0_arm64.whl", hash = "sha256:20b0f8dc160ba83b6dcc0e256846e1a02d044e13f7ea74a3d1d56ede4e48c632", size = 133433 },
+    { url = "https://files.pythonhosted.org/packages/ec/b0/b850385604334c2ce90e3ee1013bd911aedf058a934905863a6ea95e9eb4/ruamel.yaml.clib-0.2.12-cp312-cp312-manylinux2014_aarch64.whl", hash = "sha256:943f32bc9dedb3abff9879edc134901df92cfce2c3d5c9348f172f62eb2d771d", size = 647362 },
+    { url = "https://files.pythonhosted.org/packages/44/d0/3f68a86e006448fb6c005aee66565b9eb89014a70c491d70c08de597f8e4/ruamel.yaml.clib-0.2.12-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:95c3829bb364fdb8e0332c9931ecf57d9be3519241323c5274bd82f709cebc0c", size = 754118 },
+    { url = "https://files.pythonhosted.org/packages/52/a9/d39f3c5ada0a3bb2870d7db41901125dbe2434fa4f12ca8c5b83a42d7c53/ruamel.yaml.clib-0.2.12-cp312-cp312-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:749c16fcc4a2b09f28843cda5a193e0283e47454b63ec4b81eaa2242f50e4ccd", size = 706497 },
+    { url = "https://files.pythonhosted.org/packages/b0/fa/097e38135dadd9ac25aecf2a54be17ddf6e4c23e43d538492a90ab3d71c6/ruamel.yaml.clib-0.2.12-cp312-cp312-musllinux_1_1_i686.whl", hash = "sha256:bf165fef1f223beae7333275156ab2022cffe255dcc51c27f066b4370da81e31", size = 698042 },
+    { url = "https://files.pythonhosted.org/packages/ec/d5/a659ca6f503b9379b930f13bc6b130c9f176469b73b9834296822a83a132/ruamel.yaml.clib-0.2.12-cp312-cp312-musllinux_1_1_x86_64.whl", hash = "sha256:32621c177bbf782ca5a18ba4d7af0f1082a3f6e517ac2a18b3974d4edf349680", size = 745831 },
+    { url = "https://files.pythonhosted.org/packages/db/5d/36619b61ffa2429eeaefaab4f3374666adf36ad8ac6330d855848d7d36fd/ruamel.yaml.clib-0.2.12-cp312-cp312-musllinux_1_2_aarch64.whl", hash = "sha256:b82a7c94a498853aa0b272fd5bc67f29008da798d4f93a2f9f289feb8426a58d", size = 715692 },
+    { url = "https://files.pythonhosted.org/packages/b1/82/85cb92f15a4231c89b95dfe08b09eb6adca929ef7df7e17ab59902b6f589/ruamel.yaml.clib-0.2.12-cp312-cp312-win32.whl", hash = "sha256:e8c4ebfcfd57177b572e2040777b8abc537cdef58a2120e830124946aa9b42c5", size = 98777 },
+    { url = "https://files.pythonhosted.org/packages/d7/8f/c3654f6f1ddb75daf3922c3d8fc6005b1ab56671ad56ffb874d908bfa668/ruamel.yaml.clib-0.2.12-cp312-cp312-win_amd64.whl", hash = "sha256:0467c5965282c62203273b838ae77c0d29d7638c8a4e3a1c8bdd3602c10904e4", size = 115523 },
+]
+
 [[package]]
 name = "s3transfer"
 version = "0.13.0"
@@ -3098,6 +3586,68 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/69/e2/b011c38e5394c4c18fb5500778a55ec43ad6106126e74723ffaee246f56e/safetensors-0.5.3-cp38-abi3-win_amd64.whl", hash = "sha256:836cbbc320b47e80acd40e44c8682db0e8ad7123209f69b093def21ec7cafd11", size = 308878 },
 ]
 
+[[package]]
+name = "scikit-learn"
+version = "1.7.1"
+source = { registry = "https://pypi.org/simple" }
+dependencies = [
+    { name = "joblib" },
+    { name = "numpy" },
+    { name = "scipy" },
+    { name = "threadpoolctl" },
+]
+sdist = { url = "https://files.pythonhosted.org/packages/41/84/5f4af978fff619706b8961accac84780a6d298d82a8873446f72edb4ead0/scikit_learn-1.7.1.tar.gz", hash = "sha256:24b3f1e976a4665aa74ee0fcaac2b8fccc6ae77c8e07ab25da3ba6d3292b9802", size = 7190445 }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/b4/bd/a23177930abd81b96daffa30ef9c54ddbf544d3226b8788ce4c3ef1067b4/scikit_learn-1.7.1-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:90c8494ea23e24c0fb371afc474618c1019dc152ce4a10e4607e62196113851b", size = 9334838 },
+    { url = "https://files.pythonhosted.org/packages/8d/a1/d3a7628630a711e2ac0d1a482910da174b629f44e7dd8cfcd6924a4ef81a/scikit_learn-1.7.1-cp311-cp311-macosx_12_0_arm64.whl", hash = "sha256:bb870c0daf3bf3be145ec51df8ac84720d9972170786601039f024bf6d61a518", size = 8651241 },
+    { url = "https://files.pythonhosted.org/packages/26/92/85ec172418f39474c1cd0221d611345d4f433fc4ee2fc68e01f524ccc4e4/scikit_learn-1.7.1-cp311-cp311-manylinux2014_x86_64.manylinux_2_17_x86_64.whl", hash = "sha256:40daccd1b5623f39e8943ab39735cadf0bdce80e67cdca2adcb5426e987320a8", size = 9718677 },
+    { url = "https://files.pythonhosted.org/packages/df/ce/abdb1dcbb1d2b66168ec43b23ee0cee356b4cc4100ddee3943934ebf1480/scikit_learn-1.7.1-cp311-cp311-manylinux_2_27_aarch64.manylinux_2_28_aarch64.whl", hash = "sha256:30d1f413cfc0aa5a99132a554f1d80517563c34a9d3e7c118fde2d273c6fe0f7", size = 9511189 },
+    { url = "https://files.pythonhosted.org/packages/b2/3b/47b5eaee01ef2b5a80ba3f7f6ecf79587cb458690857d4777bfd77371c6f/scikit_learn-1.7.1-cp311-cp311-win_amd64.whl", hash = "sha256:c711d652829a1805a95d7fe96654604a8f16eab5a9e9ad87b3e60173415cb650", size = 8914794 },
+    { url = "https://files.pythonhosted.org/packages/cb/16/57f176585b35ed865f51b04117947fe20f130f78940c6477b6d66279c9c2/scikit_learn-1.7.1-cp312-cp312-macosx_10_13_x86_64.whl", hash = "sha256:3cee419b49b5bbae8796ecd690f97aa412ef1674410c23fc3257c6b8b85b8087", size = 9260431 },
+    { url = "https://files.pythonhosted.org/packages/67/4e/899317092f5efcab0e9bc929e3391341cec8fb0e816c4789686770024580/scikit_learn-1.7.1-cp312-cp312-macosx_12_0_arm64.whl", hash = "sha256:2fd8b8d35817b0d9ebf0b576f7d5ffbbabdb55536b0655a8aaae629d7ffd2e1f", size = 8637191 },
+    { url = "https://files.pythonhosted.org/packages/f3/1b/998312db6d361ded1dd56b457ada371a8d8d77ca2195a7d18fd8a1736f21/scikit_learn-1.7.1-cp312-cp312-manylinux2014_x86_64.manylinux_2_17_x86_64.whl", hash = "sha256:588410fa19a96a69763202f1d6b7b91d5d7a5d73be36e189bc6396bfb355bd87", size = 9486346 },
+    { url = "https://files.pythonhosted.org/packages/ad/09/a2aa0b4e644e5c4ede7006748f24e72863ba2ae71897fecfd832afea01b4/scikit_learn-1.7.1-cp312-cp312-manylinux_2_27_aarch64.manylinux_2_28_aarch64.whl", hash = "sha256:e3142f0abe1ad1d1c31a2ae987621e41f6b578144a911ff4ac94781a583adad7", size = 9290988 },
+    { url = "https://files.pythonhosted.org/packages/15/fa/c61a787e35f05f17fc10523f567677ec4eeee5f95aa4798dbbbcd9625617/scikit_learn-1.7.1-cp312-cp312-win_amd64.whl", hash = "sha256:3ddd9092c1bd469acab337d87930067c87eac6bd544f8d5027430983f1e1ae88", size = 8735568 },
+]
+
+[[package]]
+name = "scipy"
+version = "1.16.1"
+source = { registry = "https://pypi.org/simple" }
+dependencies = [
+    { name = "numpy" },
+]
+sdist = { url = "https://files.pythonhosted.org/packages/f5/4a/b927028464795439faec8eaf0b03b011005c487bb2d07409f28bf30879c4/scipy-1.16.1.tar.gz", hash = "sha256:44c76f9e8b6e8e488a586190ab38016e4ed2f8a038af7cd3defa903c0a2238b3", size = 30580861 }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/da/91/812adc6f74409b461e3a5fa97f4f74c769016919203138a3bf6fc24ba4c5/scipy-1.16.1-cp311-cp311-macosx_10_14_x86_64.whl", hash = "sha256:c033fa32bab91dc98ca59d0cf23bb876454e2bb02cbe592d5023138778f70030", size = 36552519 },
+    { url = "https://files.pythonhosted.org/packages/47/18/8e355edcf3b71418d9e9f9acd2708cc3a6c27e8f98fde0ac34b8a0b45407/scipy-1.16.1-cp311-cp311-macosx_12_0_arm64.whl", hash = "sha256:6e5c2f74e5df33479b5cd4e97a9104c511518fbd979aa9b8f6aec18b2e9ecae7", size = 28638010 },
+    { url = "https://files.pythonhosted.org/packages/d9/eb/e931853058607bdfbc11b86df19ae7a08686121c203483f62f1ecae5989c/scipy-1.16.1-cp311-cp311-macosx_14_0_arm64.whl", hash = "sha256:0a55ffe0ba0f59666e90951971a884d1ff6f4ec3275a48f472cfb64175570f77", size = 20909790 },
+    { url = "https://files.pythonhosted.org/packages/45/0c/be83a271d6e96750cd0be2e000f35ff18880a46f05ce8b5d3465dc0f7a2a/scipy-1.16.1-cp311-cp311-macosx_14_0_x86_64.whl", hash = "sha256:f8a5d6cd147acecc2603fbd382fed6c46f474cccfcf69ea32582e033fb54dcfe", size = 23513352 },
+    { url = "https://files.pythonhosted.org/packages/7c/bf/fe6eb47e74f762f933cca962db7f2c7183acfdc4483bd1c3813cfe83e538/scipy-1.16.1-cp311-cp311-manylinux2014_aarch64.manylinux_2_17_aarch64.whl", hash = "sha256:cb18899127278058bcc09e7b9966d41a5a43740b5bb8dcba401bd983f82e885b", size = 33534643 },
+    { url = "https://files.pythonhosted.org/packages/bb/ba/63f402e74875486b87ec6506a4f93f6d8a0d94d10467280f3d9d7837ce3a/scipy-1.16.1-cp311-cp311-manylinux2014_x86_64.manylinux_2_17_x86_64.whl", hash = "sha256:adccd93a2fa937a27aae826d33e3bfa5edf9aa672376a4852d23a7cd67a2e5b7", size = 35376776 },
+    { url = "https://files.pythonhosted.org/packages/c3/b4/04eb9d39ec26a1b939689102da23d505ea16cdae3dbb18ffc53d1f831044/scipy-1.16.1-cp311-cp311-musllinux_1_2_aarch64.whl", hash = "sha256:18aca1646a29ee9a0625a1be5637fa798d4d81fdf426481f06d69af828f16958", size = 35698906 },
+    { url = "https://files.pythonhosted.org/packages/04/d6/bb5468da53321baeb001f6e4e0d9049eadd175a4a497709939128556e3ec/scipy-1.16.1-cp311-cp311-musllinux_1_2_x86_64.whl", hash = "sha256:d85495cef541729a70cdddbbf3e6b903421bc1af3e8e3a9a72a06751f33b7c39", size = 38129275 },
+    { url = "https://files.pythonhosted.org/packages/c4/94/994369978509f227cba7dfb9e623254d0d5559506fe994aef4bea3ed469c/scipy-1.16.1-cp311-cp311-win_amd64.whl", hash = "sha256:226652fca853008119c03a8ce71ffe1b3f6d2844cc1686e8f9806edafae68596", size = 38644572 },
+    { url = "https://files.pythonhosted.org/packages/f8/d9/ec4864f5896232133f51382b54a08de91a9d1af7a76dfa372894026dfee2/scipy-1.16.1-cp312-cp312-macosx_10_14_x86_64.whl", hash = "sha256:81b433bbeaf35728dad619afc002db9b189e45eebe2cd676effe1fb93fef2b9c", size = 36575194 },
+    { url = "https://files.pythonhosted.org/packages/5c/6d/40e81ecfb688e9d25d34a847dca361982a6addf8e31f0957b1a54fbfa994/scipy-1.16.1-cp312-cp312-macosx_12_0_arm64.whl", hash = "sha256:886cc81fdb4c6903a3bb0464047c25a6d1016fef77bb97949817d0c0d79f9e04", size = 28594590 },
+    { url = "https://files.pythonhosted.org/packages/0e/37/9f65178edfcc629377ce9a64fc09baebea18c80a9e57ae09a52edf84880b/scipy-1.16.1-cp312-cp312-macosx_14_0_arm64.whl", hash = "sha256:15240c3aac087a522b4eaedb09f0ad061753c5eebf1ea430859e5bf8640d5919", size = 20866458 },
+    { url = "https://files.pythonhosted.org/packages/2c/7b/749a66766871ea4cb1d1ea10f27004db63023074c22abed51f22f09770e0/scipy-1.16.1-cp312-cp312-macosx_14_0_x86_64.whl", hash = "sha256:65f81a25805f3659b48126b5053d9e823d3215e4a63730b5e1671852a1705921", size = 23539318 },
+    { url = "https://files.pythonhosted.org/packages/c4/db/8d4afec60eb833a666434d4541a3151eedbf2494ea6d4d468cbe877f00cd/scipy-1.16.1-cp312-cp312-manylinux2014_aarch64.manylinux_2_17_aarch64.whl", hash = "sha256:6c62eea7f607f122069b9bad3f99489ddca1a5173bef8a0c75555d7488b6f725", size = 33292899 },
+    { url = "https://files.pythonhosted.org/packages/51/1e/79023ca3bbb13a015d7d2757ecca3b81293c663694c35d6541b4dca53e98/scipy-1.16.1-cp312-cp312-manylinux2014_x86_64.manylinux_2_17_x86_64.whl", hash = "sha256:f965bbf3235b01c776115ab18f092a95aa74c271a52577bcb0563e85738fd618", size = 35162637 },
+    { url = "https://files.pythonhosted.org/packages/b6/49/0648665f9c29fdaca4c679182eb972935b3b4f5ace41d323c32352f29816/scipy-1.16.1-cp312-cp312-musllinux_1_2_aarch64.whl", hash = "sha256:f006e323874ffd0b0b816d8c6a8e7f9a73d55ab3b8c3f72b752b226d0e3ac83d", size = 35490507 },
+    { url = "https://files.pythonhosted.org/packages/62/8f/66cbb9d6bbb18d8c658f774904f42a92078707a7c71e5347e8bf2f52bb89/scipy-1.16.1-cp312-cp312-musllinux_1_2_x86_64.whl", hash = "sha256:e8fd15fc5085ab4cca74cb91fe0a4263b1f32e4420761ddae531ad60934c2119", size = 37923998 },
+    { url = "https://files.pythonhosted.org/packages/14/c3/61f273ae550fbf1667675701112e380881905e28448c080b23b5a181df7c/scipy-1.16.1-cp312-cp312-win_amd64.whl", hash = "sha256:f7b8013c6c066609577d910d1a2a077021727af07b6fab0ee22c2f901f22352a", size = 38508060 },
+]
+
+[[package]]
+name = "semver"
+version = "3.0.4"
+source = { registry = "https://pypi.org/simple" }
+sdist = { url = "https://files.pythonhosted.org/packages/72/d1/d3159231aec234a59dd7d601e9dd9fe96f3afff15efd33c1070019b26132/semver-3.0.4.tar.gz", hash = "sha256:afc7d8c584a5ed0a11033af086e8af226a9c0b206f313e0301f8dd7b6b589602", size = 269730 }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/a6/24/4d91e05817e92e3a61c8a21e08fd0f390f5301f1c448b137c57c4bc6e543/semver-3.0.4-py3-none-any.whl", hash = "sha256:9c824d87ba7f7ab4a1890799cec8596f15c1241cb473404ea1cb0c55e4b04746", size = 17912 },
+]
+
 [[package]]
 name = "sentencepiece"
 version = "0.2.0"
@@ -3201,6 +3751,25 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/32/46/9cb0e58b2deb7f82b84065f37f3bffeb12413f947f9388e4cac22c4621ce/sortedcontainers-2.4.0-py2.py3-none-any.whl", hash = "sha256:a163dcaede0f1c021485e957a39245190e74249897e2ae4b2aa38595db237ee0", size = 29575 },
 ]
 
+[[package]]
+name = "soundfile"
+version = "0.13.1"
+source = { registry = "https://pypi.org/simple" }
+dependencies = [
+    { name = "cffi" },
+    { name = "numpy" },
+]
+sdist = { url = "https://files.pythonhosted.org/packages/e1/41/9b873a8c055582859b239be17902a85339bec6a30ad162f98c9b0288a2cc/soundfile-0.13.1.tar.gz", hash = "sha256:b2c68dab1e30297317080a5b43df57e302584c49e2942defdde0acccc53f0e5b", size = 46156 }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/64/28/e2a36573ccbcf3d57c00626a21fe51989380636e821b341d36ccca0c1c3a/soundfile-0.13.1-py2.py3-none-any.whl", hash = "sha256:a23c717560da2cf4c7b5ae1142514e0fd82d6bbd9dfc93a50423447142f2c445", size = 25751 },
+    { url = "https://files.pythonhosted.org/packages/ea/ab/73e97a5b3cc46bba7ff8650a1504348fa1863a6f9d57d7001c6b67c5f20e/soundfile-0.13.1-py2.py3-none-macosx_10_9_x86_64.whl", hash = "sha256:82dc664d19831933fe59adad199bf3945ad06d84bc111a5b4c0d3089a5b9ec33", size = 1142250 },
+    { url = "https://files.pythonhosted.org/packages/a0/e5/58fd1a8d7b26fc113af244f966ee3aecf03cb9293cb935daaddc1e455e18/soundfile-0.13.1-py2.py3-none-macosx_11_0_arm64.whl", hash = "sha256:743f12c12c4054921e15736c6be09ac26b3b3d603aef6fd69f9dde68748f2593", size = 1101406 },
+    { url = "https://files.pythonhosted.org/packages/58/ae/c0e4a53d77cf6e9a04179535766b3321b0b9ced5f70522e4caf9329f0046/soundfile-0.13.1-py2.py3-none-manylinux_2_28_aarch64.whl", hash = "sha256:9c9e855f5a4d06ce4213f31918653ab7de0c5a8d8107cd2427e44b42df547deb", size = 1235729 },
+    { url = "https://files.pythonhosted.org/packages/57/5e/70bdd9579b35003a489fc850b5047beeda26328053ebadc1fb60f320f7db/soundfile-0.13.1-py2.py3-none-manylinux_2_28_x86_64.whl", hash = "sha256:03267c4e493315294834a0870f31dbb3b28a95561b80b134f0bd3cf2d5f0e618", size = 1313646 },
+    { url = "https://files.pythonhosted.org/packages/fe/df/8c11dc4dfceda14e3003bb81a0d0edcaaf0796dd7b4f826ea3e532146bba/soundfile-0.13.1-py2.py3-none-win32.whl", hash = "sha256:c734564fab7c5ddf8e9be5bf70bab68042cd17e9c214c06e365e20d64f9a69d5", size = 899881 },
+    { url = "https://files.pythonhosted.org/packages/14/e9/6b761de83277f2f02ded7e7ea6f07828ec78e4b229b80e4ca55dd205b9dc/soundfile-0.13.1-py2.py3-none-win_amd64.whl", hash = "sha256:1e70a05a0626524a69e9f0f4dd2ec174b4e9567f4d8b6c11d38b5c289be36ee9", size = 1019162 },
+]
+
 [[package]]
 name = "soupsieve"
 version = "2.7"
@@ -3210,6 +3779,29 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/e7/9c/0e6afc12c269578be5c0c1c9f4b49a8d32770a080260c333ac04cc1c832d/soupsieve-2.7-py3-none-any.whl", hash = "sha256:6e60cc5c1ffaf1cebcc12e8188320b72071e922c2e897f737cadce79ad5d30c4", size = 36677 },
 ]
 
+[[package]]
+name = "speechbrain"
+version = "1.0.3"
+source = { registry = "https://pypi.org/simple" }
+dependencies = [
+    { name = "huggingface-hub" },
+    { name = "hyperpyyaml" },
+    { name = "joblib" },
+    { name = "numpy" },
+    { name = "packaging" },
+    { name = "scipy" },
+    { name = "sentencepiece" },
+    { name = "torch", version = "2.8.0", source = { registry = "https://download.pytorch.org/whl/cpu" }, marker = "sys_platform == 'darwin'" },
+    { name = "torch", version = "2.8.0+cpu", source = { registry = "https://download.pytorch.org/whl/cpu" }, marker = "sys_platform != 'darwin'" },
+    { name = "torchaudio", version = "2.8.0", source = { registry = "https://download.pytorch.org/whl/cpu" }, marker = "(platform_machine == 'aarch64' and sys_platform == 'linux') or sys_platform == 'darwin'" },
+    { name = "torchaudio", version = "2.8.0+cpu", source = { registry = "https://download.pytorch.org/whl/cpu" }, marker = "(platform_machine != 'aarch64' and sys_platform == 'linux') or (sys_platform != 'darwin' and sys_platform != 'linux')" },
+    { name = "tqdm" },
+]
+sdist = { url = "https://files.pythonhosted.org/packages/ab/10/87e666544a4e0cec7cbdc09f26948994831ae0f8bbc58de3bf53b68285ff/speechbrain-1.0.3.tar.gz", hash = "sha256:fcab3c6e90012cecb1eed40ea235733b550137e73da6bfa2340ba191ec714052", size = 747735 }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/58/13/e61f1085aebee17d5fc2df19fcc5177c10379be52578afbecdd615a831c9/speechbrain-1.0.3-py3-none-any.whl", hash = "sha256:9859d4c1b1fb3af3b85523c0c89f52e45a04f305622ed55f31aa32dd2fba19e9", size = 864091 },
+]
+
 [[package]]
 name = "sqlalchemy"
 version = "1.4.54"
@@ -3291,6 +3883,15 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/a2/09/77d55d46fd61b4a135c444fc97158ef34a095e5681d0a6c10b75bf356191/sympy-1.14.0-py3-none-any.whl", hash = "sha256:e091cc3e99d2141a0ba2847328f5479b05d94a6635cb96148ccb3f34671bd8f5", size = 6299353 },
 ]
 
+[[package]]
+name = "tabulate"
+version = "0.9.0"
+source = { registry = "https://pypi.org/simple" }
+sdist = { url = "https://files.pythonhosted.org/packages/ec/fe/802052aecb21e3797b8f7902564ab6ea0d60ff8ca23952079064155d1ae1/tabulate-0.9.0.tar.gz", hash = "sha256:0095b12bf5966de529c0feb1fa08671671b3368eec77d7ef7ab114be2c068b3c", size = 81090 }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/40/44/4a5f08c96eb108af5cb50b41f76142f0afa346dfa99d5296fe7202a11854/tabulate-0.9.0-py3-none-any.whl", hash = "sha256:024ca478df22e9340661486f85298cff5f6dcdba14f3813e8830015b9ed1948f", size = 35252 },
+]
+
 [[package]]
 name = "tenacity"
 version = "9.1.2"
@@ -3300,6 +3901,29 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/e5/30/643397144bfbfec6f6ef821f36f33e57d35946c44a2352d3c9f0ae847619/tenacity-9.1.2-py3-none-any.whl", hash = "sha256:f77bf36710d8b73a50b2dd155c97b870017ad21afe6ab300326b0371b3b05138", size = 28248 },
 ]
 
+[[package]]
+name = "tensorboardx"
+version = "2.6.4"
+source = { registry = "https://pypi.org/simple" }
+dependencies = [
+    { name = "numpy" },
+    { name = "packaging" },
+    { name = "protobuf" },
+]
+sdist = { url = "https://files.pythonhosted.org/packages/2b/c5/d4cc6e293fb837aaf9f76dd7745476aeba8ef7ef5146c3b3f9ee375fe7a5/tensorboardx-2.6.4.tar.gz", hash = "sha256:b163ccb7798b31100b9f5fa4d6bc22dad362d7065c2f24b51e50731adde86828", size = 4769801 }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/e0/1d/b5d63f1a6b824282b57f7b581810d20b7a28ca951f2d5b59f1eb0782c12b/tensorboardx-2.6.4-py3-none-any.whl", hash = "sha256:5970cf3a1f0a6a6e8b180ccf46f3fe832b8a25a70b86e5a237048a7c0beb18e2", size = 87201 },
+]
+
+[[package]]
+name = "threadpoolctl"
+version = "3.6.0"
+source = { registry = "https://pypi.org/simple" }
+sdist = { url = "https://files.pythonhosted.org/packages/b7/4d/08c89e34946fce2aec4fbb45c9016efd5f4d7f24af8e5d93296e935631d8/threadpoolctl-3.6.0.tar.gz", hash = "sha256:8ab8b4aa3491d812b623328249fab5302a68d2d71745c8a4c719a2fcaba9f44e", size = 21274 }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/32/d5/f9a850d79b0851d1d4ef6456097579a9005b31fea68726a4ae5f2d82ddd9/threadpoolctl-3.6.0-py3-none-any.whl", hash = "sha256:43a0b8fd5a2928500110039e43a5eed8480b918967083ea48dc3ab9f13c4a7fb", size = 18638 },
+]
+
 [[package]]
 name = "tiktoken"
 version = "0.9.0"
@@ -3440,6 +4064,40 @@ wheels = [
     { url = "https://download.pytorch.org/whl/cpu/torch-2.8.0%2Bcpu-cp312-cp312-win_arm64.whl", hash = "sha256:99fc421a5d234580e45957a7b02effbf3e1c884a5dd077afc85352c77bf41434" },
 ]
 
+[[package]]
+name = "torch-audiomentations"
+version = "0.12.0"
+source = { registry = "https://pypi.org/simple" }
+dependencies = [
+    { name = "julius" },
+    { name = "torch", version = "2.8.0", source = { registry = "https://download.pytorch.org/whl/cpu" }, marker = "sys_platform == 'darwin'" },
+    { name = "torch", version = "2.8.0+cpu", source = { registry = "https://download.pytorch.org/whl/cpu" }, marker = "sys_platform != 'darwin'" },
+    { name = "torch-pitch-shift" },
+    { name = "torchaudio", version = "2.8.0", source = { registry = "https://download.pytorch.org/whl/cpu" }, marker = "(platform_machine == 'aarch64' and sys_platform == 'linux') or sys_platform == 'darwin'" },
+    { name = "torchaudio", version = "2.8.0+cpu", source = { registry = "https://download.pytorch.org/whl/cpu" }, marker = "(platform_machine != 'aarch64' and sys_platform == 'linux') or (sys_platform != 'darwin' and sys_platform != 'linux')" },
+]
+sdist = { url = "https://files.pythonhosted.org/packages/31/8d/2f8fd7e34c75f5ee8de4310c3bd3f22270acd44d1f809e2fe7c12fbf35f8/torch_audiomentations-0.12.0.tar.gz", hash = "sha256:b02d4c5eb86376986a53eb405cca5e34f370ea9284411237508e720c529f7888", size = 52094 }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/21/9d/1ee04f49c15d2d632f6f7102061d7c07652858e6d91b58a091531034e84f/torch_audiomentations-0.12.0-py3-none-any.whl", hash = "sha256:1b80b91d2016ccf83979622cac8f702072a79b7dcc4c2bee40f00b26433a786b", size = 48506 },
+]
+
+[[package]]
+name = "torch-pitch-shift"
+version = "1.2.5"
+source = { registry = "https://pypi.org/simple" }
+dependencies = [
+    { name = "packaging" },
+    { name = "primepy" },
+    { name = "torch", version = "2.8.0", source = { registry = "https://download.pytorch.org/whl/cpu" }, marker = "sys_platform == 'darwin'" },
+    { name = "torch", version = "2.8.0+cpu", source = { registry = "https://download.pytorch.org/whl/cpu" }, marker = "sys_platform != 'darwin'" },
+    { name = "torchaudio", version = "2.8.0", source = { registry = "https://download.pytorch.org/whl/cpu" }, marker = "(platform_machine == 'aarch64' and sys_platform == 'linux') or sys_platform == 'darwin'" },
+    { name = "torchaudio", version = "2.8.0+cpu", source = { registry = "https://download.pytorch.org/whl/cpu" }, marker = "(platform_machine != 'aarch64' and sys_platform == 'linux') or (sys_platform != 'darwin' and sys_platform != 'linux')" },
+]
+sdist = { url = "https://files.pythonhosted.org/packages/79/a6/722a832bca75d5079f6731e005b3d0c2eec7c6c6863d030620952d143d57/torch_pitch_shift-1.2.5.tar.gz", hash = "sha256:6e1c7531f08d0f407a4c55e5ff8385a41355c5c5d27ab7fa08632e51defbd0ed", size = 4725 }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/27/4c/96ac2a09efb56cc3c41fb3ce9b6f4d8c0604499f7481d4a13a7b03e21382/torch_pitch_shift-1.2.5-py3-none-any.whl", hash = "sha256:6f8500cbc13f1c98b11cde1805ce5084f82cdd195c285f34287541f168a7c6a7", size = 5005 },
+]
+
 [[package]]
 name = "torchaudio"
 version = "2.8.0"
@@ -3487,6 +4145,22 @@ wheels = [
     { url = "https://download.pytorch.org/whl/cpu/torchaudio-2.8.0%2Bcpu-cp312-cp312-win_amd64.whl", hash = "sha256:9b302192b570657c1cc787a4d487ae4bbb7f2aab1c01b1fcc46757e7f86f391e" },
 ]
 
+[[package]]
+name = "torchmetrics"
+version = "1.8.2"
+source = { registry = "https://pypi.org/simple" }
+dependencies = [
+    { name = "lightning-utilities" },
+    { name = "numpy" },
+    { name = "packaging" },
+    { name = "torch", version = "2.8.0", source = { registry = "https://download.pytorch.org/whl/cpu" }, marker = "sys_platform == 'darwin'" },
+    { name = "torch", version = "2.8.0+cpu", source = { registry = "https://download.pytorch.org/whl/cpu" }, marker = "sys_platform != 'darwin'" },
+]
+sdist = { url = "https://files.pythonhosted.org/packages/85/2e/48a887a59ecc4a10ce9e8b35b3e3c5cef29d902c4eac143378526e7485cb/torchmetrics-1.8.2.tar.gz", hash = "sha256:cf64a901036bf107f17a524009eea7781c9c5315d130713aeca5747a686fe7a5", size = 580679 }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/02/21/aa0f434434c48490f91b65962b1ce863fdcce63febc166ca9fe9d706c2b6/torchmetrics-1.8.2-py3-none-any.whl", hash = "sha256:08382fd96b923e39e904c4d570f3d49e2cc71ccabd2a94e0f895d1f0dac86242", size = 983161 },
+]
+
 [[package]]
 name = "tqdm"
 version = "4.67.1"

www/app/(app)/transcripts/[transcriptId]/processing/page.tsx

@@ -11,7 +11,6 @@ import {
 import { useRouter } from "next/navigation";
 import { useTranscriptGet } from "../../../../lib/apiHooks";
 import { parseNonEmptyString } from "../../../../lib/utils";
-import { useWebSockets } from "../../useWebSockets";
 
 type TranscriptProcessing = {
   params: Promise<{
@@ -25,7 +24,6 @@ export default function TranscriptProcessing(details: TranscriptProcessing) {
   const router = useRouter();
 
   const transcript = useTranscriptGet(transcriptId);
-  useWebSockets(transcriptId);
 
   useEffect(() => {
     const status = transcript.data?.status;

www/app/(app)/transcripts/useWebRTC.ts

@@ -23,16 +23,7 @@ const useWebRTC = (
     let p: Peer;
 
     try {
-      p = new Peer({
-        initiator: true,
-        stream: stream,
-        // Disable trickle ICE: single SDP exchange (offer + answer) with all candidates.
-        // Required for HTTP-based signaling; trickle needs WebSocket for candidate exchange.
-        trickle: false,
-        config: {
-          iceServers: [{ urls: "stun:stun.l.google.com:19302" }],
-        },
-      });
+      p = new Peer({ initiator: true, stream: stream });
     } catch (error) {
       setError(error as Error, "Error creating WebRTC");
       return;

www/app/(app)/transcripts/useWebSockets.ts

@@ -431,7 +431,6 @@ export const useWebSockets = (transcriptId: string | null): UseWebSockets => {
               );
             }
             setStatus(message.data);
-            invalidateTranscript(queryClient, transcriptId as NonEmptyString);
             if (message.data.value === "ended") {
               ws.close();
             }

www/app/[roomName]/components/DailyRoom.tsx

@@ -24,15 +24,24 @@ import { useAuth } from "../../lib/AuthProvider";
 import { useConsentDialog } from "../../lib/consent";
 import {
   useRoomJoinMeeting,
+  useRoomJoinedMeeting,
+  useRoomLeaveMeeting,
   useMeetingStartRecording,
+  leaveRoomPostUrl,
+  LeaveRoomBody,
 } from "../../lib/apiHooks";
 import { omit } from "remeda";
 import {
   assertExists,
+  assertExistsAndNonEmptyString,
   NonEmptyString,
   parseNonEmptyString,
 } from "../../lib/utils";
-import { assertMeetingId, DailyRecordingType } from "../../lib/types";
+import {
+  assertMeetingId,
+  DailyRecordingType,
+  MeetingId,
+} from "../../lib/types";
 import { useUuidV5 } from "react-uuid-hook";
 
 const CONSENT_BUTTON_ID = "recording-consent";
@@ -179,6 +188,58 @@ const useFrame = (
   ] as const;
 };
 
+const leaveDaily = () => {
+  const frame = DailyIframe.getCallInstance();
+  frame?.leave();
+};
+
+const useDirtyDisconnects = (
+  meetingId: NonEmptyString,
+  roomName: NonEmptyString,
+) => {
+  useEffect(() => {
+    if (!meetingId || !roomName) return;
+
+    const handleBeforeUnload = () => {
+      leaveDaily();
+      navigator.sendBeacon(
+        leaveRoomPostUrl(
+          {
+            room_name: roomName,
+            meeting_id: meetingId,
+          },
+          {
+            delay_seconds: 5,
+          },
+        ),
+        undefined satisfies LeaveRoomBody,
+      );
+    };
+    window.addEventListener("beforeunload", handleBeforeUnload);
+    return () => window.removeEventListener("beforeunload", handleBeforeUnload);
+  }, [meetingId, roomName]);
+};
+
+const useDisconnects = (
+  meetingId: NonEmptyString,
+  roomName: NonEmptyString,
+  leaveMutation: ReturnType<typeof useRoomLeaveMeeting>,
+) => {
+  useDirtyDisconnects(meetingId, roomName);
+
+  useEffect(() => {
+    return () => {
+      leaveDaily();
+      leaveMutation.mutate({
+        params: {
+          path: { meeting_id: meetingId, room_name: roomName },
+          query: { delay_seconds: 5 },
+        },
+      });
+    };
+  }, [meetingId, roomName]);
+};
+
 export default function DailyRoom({ meeting, room }: DailyRoomProps) {
   const router = useRouter();
   const params = useParams();
@@ -186,6 +247,8 @@ export default function DailyRoom({ meeting, room }: DailyRoomProps) {
   const authLastUserId = auth.lastUserId;
   const [container, setContainer] = useState<HTMLDivElement | null>(null);
   const joinMutation = useRoomJoinMeeting();
+  const joinedMutation = useRoomJoinedMeeting();
+  const leaveMutation = useRoomLeaveMeeting();
   const startRecordingMutation = useMeetingStartRecording();
   const [joinedMeeting, setJoinedMeeting] = useState<Meeting | null>(null);
 
@@ -195,7 +258,9 @@ export default function DailyRoom({ meeting, room }: DailyRoomProps) {
     useUuidV5(meeting.id, RAW_TRACKS_NAMESPACE)[0],
   );
 
-  const roomName = params?.roomName as string;
+  if (typeof params.roomName === "object")
+    throw new Error(`Invalid room name in params. array? ${params.roomName}`);
+  const roomName = assertExistsAndNonEmptyString(params.roomName);
 
   const {
     showConsentModal,
@@ -237,6 +302,8 @@ export default function DailyRoom({ meeting, room }: DailyRoomProps) {
     router.push("/browse");
   }, [router]);
 
+  useDisconnects(meeting.id as MeetingId, roomName, leaveMutation);
+
   const handleCustomButtonClick = useCallback(
     (ev: DailyEventObjectCustomButtonClick) => {
       if (ev.button_id === CONSENT_BUTTON_ID) {
@@ -249,6 +316,15 @@ export default function DailyRoom({ meeting, room }: DailyRoomProps) {
   );
 
   const handleFrameJoinMeeting = useCallback(() => {
+    joinedMutation.mutate({
+      params: {
+        path: {
+          room_name: roomName,
+          meeting_id: meeting.id,
+        },
+      },
+    });
+
     if (meeting.recording_type === "cloud") {
       console.log("Starting dual recording via REST API", {
         cloudInstanceId,
@@ -308,8 +384,10 @@ export default function DailyRoom({ meeting, room }: DailyRoomProps) {
       startRecordingWithRetry("raw-tracks", rawTracksInstanceId);
     }
   }, [
-    meeting.recording_type,
+    joinedMutation,
+    roomName,
     meeting.id,
+    meeting.recording_type,
     startRecordingMutation,
     cloudInstanceId,
     rawTracksInstanceId,

www/app/lib/apiHooks.ts

@@ -1,12 +1,13 @@
 "use client";
 
-import { $api } from "./apiClient";
+import { $api, API_URL } from "./apiClient";
 import { useError } from "../(errors)/errorContext";
 import { QueryClient, useQueryClient } from "@tanstack/react-query";
-import type { components } from "../reflector-api";
+import type { components, operations } from "../reflector-api";
 import { useAuth } from "./AuthProvider";
 import { MeetingId } from "./types";
 import { NonEmptyString } from "./utils";
+import { createFinalURL, createQuerySerializer } from "openapi-fetch";
 
 /*
  * XXX error types returned from the hooks are not always correct; declared types are ValidationError but real type could be string or any other
@@ -807,6 +808,44 @@ export function useRoomJoinMeeting() {
   );
 }
 
+export const LEAVE_ROOM_POST_URL_TEMPLATE =
+  "/v1/rooms/{room_name}/meetings/{meeting_id}/leave" as const;
+
+export const leaveRoomPostUrl = (
+  path: operations["v1_rooms_leave_meeting"]["parameters"]["path"],
+  query?: operations["v1_rooms_leave_meeting"]["parameters"]["query"],
+): string =>
+  createFinalURL(LEAVE_ROOM_POST_URL_TEMPLATE, {
+    baseUrl: API_URL,
+    params: { path, query },
+    querySerializer: createQuerySerializer(),
+  });
+
+export type LeaveRoomBody = operations["v1_rooms_leave_meeting"]["requestBody"];
+
+export function useRoomLeaveMeeting() {
+  return $api.useMutation("post", LEAVE_ROOM_POST_URL_TEMPLATE);
+}
+
+export const JOINED_ROOM_POST_URL_TEMPLATE =
+  "/v1/rooms/{room_name}/meetings/{meeting_id}/joined" as const;
+
+export const joinedRoomPostUrl = (
+  params: operations["v1_rooms_joined_meeting"]["parameters"]["path"],
+): string =>
+  createFinalURL(JOINED_ROOM_POST_URL_TEMPLATE, {
+    baseUrl: API_URL,
+    params: { path: params },
+    querySerializer: () => "",
+  });
+
+export type JoinedRoomBody =
+  operations["v1_rooms_joined_meeting"]["requestBody"];
+
+export function useRoomJoinedMeeting() {
+  return $api.useMutation("post", JOINED_ROOM_POST_URL_TEMPLATE);
+}
+
 export function useRoomIcsSync() {
   const { setError } = useError();
 

www/app/reflector-api.d.ts

@@ -171,6 +171,48 @@ export interface paths {
     patch?: never;
     trace?: never;
   };
+  "/v1/rooms/{room_name}/meetings/{meeting_id}/joined": {
+    parameters: {
+      query?: never;
+      header?: never;
+      path?: never;
+      cookie?: never;
+    };
+    get?: never;
+    put?: never;
+    /**
+     * Rooms Joined Meeting
+     * @description Trigger presence poll (ideally when user actually joins meeting in Daily iframe)
+     */
+    post: operations["v1_rooms_joined_meeting"];
+    delete?: never;
+    options?: never;
+    head?: never;
+    patch?: never;
+    trace?: never;
+  };
+  "/v1/rooms/{room_name}/meetings/{meeting_id}/leave": {
+    parameters: {
+      query?: never;
+      header?: never;
+      path?: never;
+      cookie?: never;
+    };
+    get?: never;
+    put?: never;
+    /**
+     * Rooms Leave Meeting
+     * @description Trigger presence recheck when user leaves meeting (e.g., tab close/navigation).
+     *
+     *     Queues presence poll with optional delay to allow Daily.co to detect disconnect.
+     */
+    post: operations["v1_rooms_leave_meeting"];
+    delete?: never;
+    options?: never;
+    head?: never;
+    patch?: never;
+    trace?: never;
+  };
   "/v1/rooms/{room_id}/webhook/test": {
     parameters: {
       query?: never;
@@ -2435,6 +2477,72 @@ export interface operations {
       };
     };
   };
+  v1_rooms_joined_meeting: {
+    parameters: {
+      query?: never;
+      header?: never;
+      path: {
+        room_name: string;
+        meeting_id: string;
+      };
+      cookie?: never;
+    };
+    requestBody?: never;
+    responses: {
+      /** @description Successful Response */
+      200: {
+        headers: {
+          [name: string]: unknown;
+        };
+        content: {
+          "application/json": unknown;
+        };
+      };
+      /** @description Validation Error */
+      422: {
+        headers: {
+          [name: string]: unknown;
+        };
+        content: {
+          "application/json": components["schemas"]["HTTPValidationError"];
+        };
+      };
+    };
+  };
+  v1_rooms_leave_meeting: {
+    parameters: {
+      query?: {
+        delay_seconds?: number;
+      };
+      header?: never;
+      path: {
+        room_name: string;
+        meeting_id: string;
+      };
+      cookie?: never;
+    };
+    requestBody?: never;
+    responses: {
+      /** @description Successful Response */
+      200: {
+        headers: {
+          [name: string]: unknown;
+        };
+        content: {
+          "application/json": unknown;
+        };
+      };
+      /** @description Validation Error */
+      422: {
+        headers: {
+          [name: string]: unknown;
+        };
+        content: {
+          "application/json": components["schemas"]["HTTPValidationError"];
+        };
+      };
+    };
+  };
   v1_rooms_test_webhook: {
     parameters: {
       query?: never;