chore(main): release 0.32.1 (#840)

* Add Modal backend for audio padding

- Create reflector_padding.py Modal deployment (CPU-based)
- Add PaddingWorkflow with conditional Modal/local backend
- Update deploy-all.sh to include padding deployment

---------

Co-authored-by: Igor Loskutov <igor.loskutoff@gmail.com>

.gitignore

@@ -1,5 +1,6 @@
 .DS_Store
 server/.env
+server/.env.production
 .env
 Caddyfile
 server/exportdanswer

.gitleaksignore

@@ -4,4 +4,3 @@ 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/reflector/worker/process.py:generic-api-key:594

CHANGELOG.md

@@ -1,5 +1,20 @@
 # Changelog
 
+## [0.32.1](https://github.com/Monadical-SAS/reflector/compare/v0.32.0...v0.32.1) (2026-01-30)
+
+
+### Bug Fixes
+
+* daily multitrack pipeline finalze dependency fix ([23eb137](https://github.com/Monadical-SAS/reflector/commit/23eb1371cb9348c4b81eb12ad506b582f8a4799e))
+* match httpx pad with hatchet audio timeout ([c05d1f0](https://github.com/Monadical-SAS/reflector/commit/c05d1f03cd8369fc06efd455527e50246887efd0))
+
+## [0.32.0](https://github.com/Monadical-SAS/reflector/compare/v0.31.0...v0.32.0) (2026-01-30)
+
+
+### Features
+
+* modal padding ([#837](https://github.com/Monadical-SAS/reflector/issues/837)) ([7fde64e](https://github.com/Monadical-SAS/reflector/commit/7fde64e2529a1d37b0f7507c62d983a7bd0b5b89))
+
 ## [0.31.0](https://github.com/Monadical-SAS/reflector/compare/v0.30.0...v0.31.0) (2026-01-23)
 
 

gpu/modal_deployments/deploy-all.sh

@@ -131,6 +131,15 @@ if [ -z "$DIARIZER_URL" ]; then
 fi
 echo "  -> $DIARIZER_URL"
 
+echo ""
+echo "Deploying padding (CPU audio processing via Modal SDK)..."
+modal deploy reflector_padding.py
+if [ $? -ne 0 ]; then
+    echo "Error: Failed to deploy padding. Check Modal dashboard for details."
+    exit 1
+fi
+echo "  -> reflector-padding.pad_track (Modal SDK function)"
+
 # --- Output Configuration ---
 echo ""
 echo "=========================================="
@@ -147,4 +156,6 @@ echo ""
 echo "DIARIZATION_BACKEND=modal"
 echo "DIARIZATION_URL=$DIARIZER_URL"
 echo "DIARIZATION_MODAL_API_KEY=$API_KEY"
+echo ""
+echo "# Padding uses Modal SDK (requires MODAL_TOKEN_ID/SECRET in worker containers)"
 echo "# --- End Modal Configuration ---"

gpu/modal_deployments/reflector_padding.py

@@ -0,0 +1,277 @@
+"""
+Reflector GPU backend - audio padding
+======================================
+
+CPU-intensive audio padding service for adding silence to audio tracks.
+Uses PyAV filter graph (adelay) for precise track synchronization.
+
+IMPORTANT: This padding logic is duplicated from server/reflector/utils/audio_padding.py
+for Modal deployment isolation (Modal can't import from server/reflector/). If you modify
+the PyAV filter graph or padding algorithm, you MUST update both:
+  - gpu/modal_deployments/reflector_padding.py (this file)
+  - server/reflector/utils/audio_padding.py
+
+Constants duplicated from server/reflector/utils/audio_constants.py for same reason.
+"""
+
+import os
+import tempfile
+from fractions import Fraction
+import math
+import asyncio
+
+import modal
+
+S3_TIMEOUT = 60 # happens 2 times
+PADDING_TIMEOUT = 600 + (S3_TIMEOUT * 2)
+SCALEDOWN_WINDOW = 60  # The maximum duration (in seconds) that individual containers can remain idle when scaling down.
+DISCONNECT_CHECK_INTERVAL = 2  # Check for client disconnect
+
+
+app = modal.App("reflector-padding")
+
+# CPU-based image
+image = (
+    modal.Image.debian_slim(python_version="3.12")
+    .apt_install("ffmpeg")  # Required by PyAV
+    .pip_install(
+        "av==13.1.0",  # PyAV for audio processing
+        "requests==2.32.3",  # HTTP for presigned URL downloads/uploads
+        "fastapi==0.115.12",  # API framework
+    )
+)
+
+# ref B0F71CE8-FC59-4AA5-8414-DAFB836DB711
+OPUS_STANDARD_SAMPLE_RATE = 48000
+# ref B0F71CE8-FC59-4AA5-8414-DAFB836DB711
+OPUS_DEFAULT_BIT_RATE = 128000
+
+
+@app.function(
+    cpu=2.0,
+    timeout=PADDING_TIMEOUT,
+    scaledown_window=SCALEDOWN_WINDOW,
+    image=image,
+)
+@modal.asgi_app()
+def web():
+    from fastapi import FastAPI, Request, HTTPException
+    from pydantic import BaseModel
+
+    class PaddingRequest(BaseModel):
+        track_url: str
+        output_url: str
+        start_time_seconds: float
+        track_index: int
+
+    class PaddingResponse(BaseModel):
+        size: int
+        cancelled: bool = False
+
+    web_app = FastAPI()
+
+    @web_app.post("/pad")
+    async def pad_track_endpoint(request: Request, req: PaddingRequest) -> PaddingResponse:
+        """Modal web endpoint for padding audio tracks with disconnect detection.
+        """
+        import logging
+
+        logging.basicConfig(level=logging.INFO, format="%(asctime)s - %(levelname)s - %(message)s")
+        logger = logging.getLogger(__name__)
+
+        if not req.track_url:
+            raise HTTPException(status_code=400, detail="track_url cannot be empty")
+        if not req.output_url:
+            raise HTTPException(status_code=400, detail="output_url cannot be empty")
+        if req.start_time_seconds <= 0:
+            raise HTTPException(status_code=400, detail=f"start_time_seconds must be positive, got {req.start_time_seconds}")
+        if req.start_time_seconds > 18000:
+            raise HTTPException(status_code=400, detail=f"start_time_seconds exceeds maximum 18000s (5 hours)")
+
+        logger.info(f"Padding request: track {req.track_index}, delay={req.start_time_seconds}s")
+
+        # Thread-safe cancellation flag shared between async disconnect checker and blocking thread
+        import threading
+        cancelled = threading.Event()
+
+        async def check_disconnect():
+            """Background task to check for client disconnect every 2 seconds."""
+            while not cancelled.is_set():
+                await asyncio.sleep(DISCONNECT_CHECK_INTERVAL)
+                if await request.is_disconnected():
+                    logger.warning("Client disconnected, setting cancellation flag")
+                    cancelled.set()
+                    break
+
+        # Start disconnect checker in background
+        disconnect_task = asyncio.create_task(check_disconnect())
+
+        try:
+            result = await asyncio.get_event_loop().run_in_executor(
+                None, _pad_track_blocking, req, cancelled, logger
+            )
+            return PaddingResponse(**result)
+        finally:
+            cancelled.set()
+            disconnect_task.cancel()
+            try:
+                await disconnect_task
+            except asyncio.CancelledError:
+                pass
+
+    def _pad_track_blocking(req, cancelled, logger) -> dict:
+        """Blocking CPU-bound padding work with periodic cancellation checks.
+
+        Args:
+            cancelled: threading.Event for thread-safe cancellation signaling
+        """
+        import av
+        import requests
+        from av.audio.resampler import AudioResampler
+        import time
+
+        temp_dir = tempfile.mkdtemp()
+        input_path = None
+        output_path = None
+        last_check = time.time()
+
+        try:
+            logger.info("Downloading track for padding")
+            response = requests.get(req.track_url, stream=True, timeout=S3_TIMEOUT)
+            response.raise_for_status()
+
+            input_path = os.path.join(temp_dir, "track.webm")
+            total_bytes = 0
+            chunk_count = 0
+            with open(input_path, "wb") as f:
+                for chunk in response.iter_content(chunk_size=8192):
+                    if chunk:
+                        f.write(chunk)
+                        total_bytes += len(chunk)
+                        chunk_count += 1
+
+                        # Check for cancellation every arbitrary amount of chunks
+                        if chunk_count % 12 == 0:
+                            now = time.time()
+                            if now - last_check >= DISCONNECT_CHECK_INTERVAL:
+                                if cancelled.is_set():
+                                    logger.info("Cancelled during download, exiting early")
+                                    return {"size": 0, "cancelled": True}
+                                last_check = now
+            logger.info(f"Track downloaded: {total_bytes} bytes")
+
+            if cancelled.is_set():
+                logger.info("Cancelled after download, exiting early")
+                return {"size": 0, "cancelled": True}
+
+            # Apply padding using PyAV
+            output_path = os.path.join(temp_dir, "padded.webm")
+            delay_ms = math.floor(req.start_time_seconds * 1000)
+            logger.info(f"Padding track {req.track_index} with {delay_ms}ms delay using PyAV")
+
+            in_container = av.open(input_path)
+            in_stream = next((s for s in in_container.streams if s.type == "audio"), None)
+            if in_stream is None:
+                raise ValueError("No audio stream in input")
+
+            with av.open(output_path, "w", format="webm") as out_container:
+                out_stream = out_container.add_stream("libopus", rate=OPUS_STANDARD_SAMPLE_RATE)
+                out_stream.bit_rate = OPUS_DEFAULT_BIT_RATE
+                graph = av.filter.Graph()
+
+                abuf_args = (
+                    f"time_base=1/{OPUS_STANDARD_SAMPLE_RATE}:"
+                    f"sample_rate={OPUS_STANDARD_SAMPLE_RATE}:"
+                    f"sample_fmt=s16:"
+                    f"channel_layout=stereo"
+                )
+                src = graph.add("abuffer", args=abuf_args, name="src")
+                aresample_f = graph.add("aresample", args="async=1", name="ares")
+                delays_arg = f"{delay_ms}|{delay_ms}"
+                adelay_f = graph.add("adelay", args=f"delays={delays_arg}:all=1", name="delay")
+                sink = graph.add("abuffersink", name="sink")
+
+                src.link_to(aresample_f)
+                aresample_f.link_to(adelay_f)
+                adelay_f.link_to(sink)
+                graph.configure()
+
+                resampler = AudioResampler(
+                    format="s16", layout="stereo", rate=OPUS_STANDARD_SAMPLE_RATE
+                )
+
+                for frame in in_container.decode(in_stream):
+                    # Check for cancellation periodically
+                    now = time.time()
+                    if now - last_check >= DISCONNECT_CHECK_INTERVAL:
+                        if cancelled.is_set():
+                            logger.info("Cancelled during processing, exiting early")
+                            in_container.close()
+                            return {"size": 0, "cancelled": True}
+                        last_check = now
+
+                    out_frames = resampler.resample(frame) or []
+                    for rframe in out_frames:
+                        rframe.sample_rate = OPUS_STANDARD_SAMPLE_RATE
+                        rframe.time_base = Fraction(1, OPUS_STANDARD_SAMPLE_RATE)
+                        src.push(rframe)
+
+                        while True:
+                            try:
+                                f_out = sink.pull()
+                            except Exception:
+                                break
+                            f_out.sample_rate = OPUS_STANDARD_SAMPLE_RATE
+                            f_out.time_base = Fraction(1, OPUS_STANDARD_SAMPLE_RATE)
+                            for packet in out_stream.encode(f_out):
+                                out_container.mux(packet)
+
+                # Flush filter graph
+                src.push(None)
+                while True:
+                    try:
+                        f_out = sink.pull()
+                    except Exception:
+                        break
+                    f_out.sample_rate = OPUS_STANDARD_SAMPLE_RATE
+                    f_out.time_base = Fraction(1, OPUS_STANDARD_SAMPLE_RATE)
+                    for packet in out_stream.encode(f_out):
+                        out_container.mux(packet)
+
+                # Flush encoder
+                for packet in out_stream.encode(None):
+                    out_container.mux(packet)
+
+            in_container.close()
+
+            file_size = os.path.getsize(output_path)
+            logger.info(f"Padding complete: {file_size} bytes")
+
+            logger.info("Uploading padded track to S3")
+
+            with open(output_path, "rb") as f:
+                upload_response = requests.put(req.output_url, data=f, timeout=S3_TIMEOUT)
+
+            upload_response.raise_for_status()
+            logger.info(f"Upload complete: {file_size} bytes")
+
+            return {"size": file_size}
+
+        finally:
+            if input_path and os.path.exists(input_path):
+                try:
+                    os.unlink(input_path)
+                except Exception as e:
+                    logger.warning(f"Failed to cleanup input file: {e}")
+            if output_path and os.path.exists(output_path):
+                try:
+                    os.unlink(output_path)
+                except Exception as e:
+                    logger.warning(f"Failed to cleanup output file: {e}")
+            try:
+                os.rmdir(temp_dir)
+            except Exception as e:
+                logger.warning(f"Failed to cleanup temp directory: {e}")
+
+    return web_app
+

server/pyproject.toml

@@ -8,7 +8,7 @@ readme = "README.md"
 dependencies = [
     "aiohttp>=3.9.0",
     "aiohttp-cors>=0.7.0",
-    "av>=10.0.0",
+    "av>=15.0.0",
     "requests>=2.31.0",
     "aiortc>=1.5.0",
     "sortedcontainers>=2.4.0",

server/reflector/hatchet/constants.py

@@ -8,8 +8,7 @@ from enum import StrEnum
 class TaskName(StrEnum):
     GET_RECORDING = "get_recording"
     GET_PARTICIPANTS = "get_participants"
-    PROCESS_PADDINGS = "process_paddings"
+    PROCESS_TRACKS = "process_tracks"
-    PROCESS_TRANSCRIPTIONS = "process_transcriptions"
     MIXDOWN_TRACKS = "mixdown_tracks"
     GENERATE_WAVEFORM = "generate_waveform"
     DETECT_TOPICS = "detect_topics"
@@ -38,5 +37,5 @@ LLM_RATE_LIMIT_PER_SECOND = 10
 TIMEOUT_SHORT = 60  # Quick operations: API calls, DB updates
 TIMEOUT_MEDIUM = 120  # Single LLM calls, waveform generation
 TIMEOUT_LONG = 180  # Action items (larger context LLM)
-TIMEOUT_AUDIO = 300  # Audio processing: padding, mixdown
+TIMEOUT_AUDIO = 720  # Audio processing: padding, mixdown
 TIMEOUT_HEAVY = 600  # Transcription, fan-out LLM tasks

server/reflector/hatchet/run_workers_cpu.py

@@ -1,9 +1,9 @@
 """
 CPU-heavy worker pool for audio processing tasks.
-Handles: mixdown_tracks only (serialized with max_runs=1)
+Handles ONLY: mixdown_tracks
 
 Configuration:
-- slots=1: Only one mixdown at a time
+- slots=1: Only mixdown (already serialized globally with max_runs=1)
 - Worker affinity: pool=cpu-heavy
 """
 
@@ -26,7 +26,7 @@ def main():
 
     cpu_worker = hatchet.worker(
         "cpu-worker-pool",
-        slots=1,
+        slots=1,  # Only 1 mixdown at a time (already serialized globally)
         labels={
             "pool": "cpu-heavy",
         },

server/reflector/hatchet/run_workers_llm.py

@@ -1,16 +1,15 @@
 """
 LLM/I/O worker pool for all non-CPU tasks.
-Handles: all tasks except mixdown_tracks (padding, transcription, LLM inference, orchestration)
+Handles: all tasks except mixdown_tracks (transcription, LLM inference, orchestration)
 """
 
 from reflector.hatchet.client import HatchetClientManager
 from reflector.hatchet.workflows.daily_multitrack_pipeline import (
     daily_multitrack_pipeline,
 )
-from reflector.hatchet.workflows.padding_workflow import padding_workflow
 from reflector.hatchet.workflows.subject_processing import subject_workflow
 from reflector.hatchet.workflows.topic_chunk_processing import topic_chunk_workflow
-from reflector.hatchet.workflows.transcription_workflow import transcription_workflow
+from reflector.hatchet.workflows.track_processing import track_workflow
 from reflector.logger import logger
 
 SLOTS = 10
@@ -30,7 +29,7 @@ def main():
 
     llm_worker = hatchet.worker(
         WORKER_NAME,
-        slots=SLOTS,
+        slots=SLOTS,  # not all slots are probably used
         labels={
             "pool": POOL,
         },
@@ -38,8 +37,7 @@ def main():
             daily_multitrack_pipeline,
             topic_chunk_workflow,
             subject_workflow,
-            padding_workflow,
+            track_workflow,
-            transcription_workflow,
         ],
     )
 

server/reflector/hatchet/workflows/__init__.py

@@ -4,10 +4,6 @@ from reflector.hatchet.workflows.daily_multitrack_pipeline import (
     PipelineInput,
     daily_multitrack_pipeline,
 )
-from reflector.hatchet.workflows.padding_workflow import (
-    PaddingInput,
-    padding_workflow,
-)
 from reflector.hatchet.workflows.subject_processing import (
     SubjectInput,
     subject_workflow,
@@ -16,20 +12,15 @@ from reflector.hatchet.workflows.topic_chunk_processing import (
     TopicChunkInput,
     topic_chunk_workflow,
 )
-from reflector.hatchet.workflows.transcription_workflow import (
+from reflector.hatchet.workflows.track_processing import TrackInput, track_workflow
-    TranscriptionInput,
-    transcription_workflow,
-)
 
 __all__ = [
     "daily_multitrack_pipeline",
     "subject_workflow",
     "topic_chunk_workflow",
-    "padding_workflow",
+    "track_workflow",
-    "transcription_workflow",
     "PipelineInput",
     "SubjectInput",
     "TopicChunkInput",
-    "PaddingInput",
+    "TrackInput",
-    "TranscriptionInput",
 ]

server/reflector/hatchet/workflows/daily_multitrack_pipeline.py

@@ -54,9 +54,8 @@ from reflector.hatchet.workflows.models import (
     PadTrackResult,
     ParticipantInfo,
     ParticipantsResult,
-    ProcessPaddingsResult,
     ProcessSubjectsResult,
-    ProcessTranscriptionsResult,
+    ProcessTracksResult,
     RecapResult,
     RecordingResult,
     SubjectsResult,
@@ -69,7 +68,6 @@ from reflector.hatchet.workflows.models import (
     WebhookResult,
     ZulipResult,
 )
-from reflector.hatchet.workflows.padding_workflow import PaddingInput, padding_workflow
 from reflector.hatchet.workflows.subject_processing import (
     SubjectInput,
     subject_workflow,
@@ -78,10 +76,7 @@ from reflector.hatchet.workflows.topic_chunk_processing import (
     TopicChunkInput,
     topic_chunk_workflow,
 )
-from reflector.hatchet.workflows.transcription_workflow import (
+from reflector.hatchet.workflows.track_processing import TrackInput, track_workflow
-    TranscriptionInput,
-    transcription_workflow,
-)
 from reflector.logger import logger
 from reflector.pipelines import topic_processing
 from reflector.processors import AudioFileWriterProcessor
@@ -409,115 +404,72 @@ async def get_participants(input: PipelineInput, ctx: Context) -> ParticipantsRe
     execution_timeout=timedelta(seconds=TIMEOUT_HEAVY),
     retries=3,
 )
-@with_error_handling(TaskName.PROCESS_PADDINGS)
+@with_error_handling(TaskName.PROCESS_TRACKS)
-async def process_paddings(input: PipelineInput, ctx: Context) -> ProcessPaddingsResult:
+async def process_tracks(input: PipelineInput, ctx: Context) -> ProcessTracksResult:
-    """Spawn child workflows for each track to apply padding (dynamic fan-out)."""
+    """Spawn child workflows for each track (dynamic fan-out)."""
-    ctx.log(f"process_paddings: spawning {len(input.tracks)} padding workflows")
+    ctx.log(f"process_tracks: spawning {len(input.tracks)} track workflows")
+
+    participants_result = ctx.task_output(get_participants)
+    source_language = participants_result.source_language
 
     bulk_runs = [
-        padding_workflow.create_bulk_run_item(
+        track_workflow.create_bulk_run_item(
-            input=PaddingInput(
+            input=TrackInput(
                 track_index=i,
                 s3_key=track["s3_key"],
                 bucket_name=input.bucket_name,
                 transcript_id=input.transcript_id,
+                language=source_language,
             )
         )
         for i, track in enumerate(input.tracks)
     ]
 
-    results = await padding_workflow.aio_run_many(bulk_runs)
+    results = await track_workflow.aio_run_many(bulk_runs)
-
-    padded_tracks = []
-    created_padded_files = []
-
-    for result in results:
-        pad_result = PadTrackResult(**result[TaskName.PAD_TRACK])
-
-        padded_tracks.append(
-            PaddedTrackInfo(
-                key=pad_result.padded_key,
-                bucket_name=pad_result.bucket_name,
-                track_index=pad_result.track_index,
-            )
-        )
-
-        if pad_result.size > 0:
-            storage_path = f"file_pipeline_hatchet/{input.transcript_id}/tracks/padded_{pad_result.track_index}.webm"
-            created_padded_files.append(storage_path)
-
-    ctx.log(f"process_paddings complete: {len(padded_tracks)} padded tracks")
-
-    return ProcessPaddingsResult(
-        padded_tracks=padded_tracks,
-        num_tracks=len(input.tracks),
-        created_padded_files=list(created_padded_files),
-    )
-
-
-@daily_multitrack_pipeline.task(
-    parents=[process_paddings],
-    execution_timeout=timedelta(seconds=TIMEOUT_HEAVY),
-    retries=3,
-)
-@with_error_handling(TaskName.PROCESS_TRANSCRIPTIONS)
-async def process_transcriptions(
-    input: PipelineInput, ctx: Context
-) -> ProcessTranscriptionsResult:
-    """Spawn child workflows for each padded track to transcribe (dynamic fan-out)."""
-    participants_result = ctx.task_output(get_participants)
-    paddings_result = ctx.task_output(process_paddings)
-
-    source_language = participants_result.source_language
-    if not source_language:
-        raise ValueError("source_language is required for transcription")
 
     target_language = participants_result.target_language
-    padded_tracks = paddings_result.padded_tracks
-
-    if not padded_tracks:
-        raise ValueError("No padded tracks available for transcription")
-
-    ctx.log(
-        f"process_transcriptions: spawning {len(padded_tracks)} transcription workflows"
-    )
-
-    bulk_runs = [
-        transcription_workflow.create_bulk_run_item(
-            input=TranscriptionInput(
-                track_index=padded_track.track_index,
-                padded_key=padded_track.key,
-                bucket_name=padded_track.bucket_name,
-                language=source_language,
-            )
-        )
-        for padded_track in padded_tracks
-    ]
-
-    results = await transcription_workflow.aio_run_many(bulk_runs)
 
     track_words: list[list[Word]] = []
+    padded_tracks = []
+    created_padded_files = set()
+
     for result in results:
         transcribe_result = TranscribeTrackResult(**result[TaskName.TRANSCRIBE_TRACK])
         track_words.append(transcribe_result.words)
 
+        pad_result = PadTrackResult(**result[TaskName.PAD_TRACK])
+
+        # Store S3 key info (not presigned URL) - consumer tasks presign on demand
+        if pad_result.padded_key:
+            padded_tracks.append(
+                PaddedTrackInfo(
+                    key=pad_result.padded_key, bucket_name=pad_result.bucket_name
+                )
+            )
+
+        if pad_result.size > 0:
+            storage_path = f"file_pipeline_hatchet/{input.transcript_id}/tracks/padded_{pad_result.track_index}.webm"
+            created_padded_files.add(storage_path)
+
     all_words = [word for words in track_words for word in words]
     all_words.sort(key=lambda w: w.start)
 
     ctx.log(
-        f"process_transcriptions complete: {len(all_words)} words from {len(padded_tracks)} tracks"
+        f"process_tracks complete: {len(all_words)} words from {len(input.tracks)} tracks"
     )
 
-    return ProcessTranscriptionsResult(
+    return ProcessTracksResult(
         all_words=all_words,
+        padded_tracks=padded_tracks,
         word_count=len(all_words),
         num_tracks=len(input.tracks),
         target_language=target_language,
+        created_padded_files=list(created_padded_files),
     )
 
 
 @daily_multitrack_pipeline.task(
-    parents=[process_paddings],
+    parents=[process_tracks],
     execution_timeout=timedelta(seconds=TIMEOUT_AUDIO),
     retries=3,
     desired_worker_labels={
@@ -537,12 +489,12 @@ async def process_transcriptions(
 )
 @with_error_handling(TaskName.MIXDOWN_TRACKS)
 async def mixdown_tracks(input: PipelineInput, ctx: Context) -> MixdownResult:
-    """Mix all padded tracks into single audio file using PyAV."""
+    """Mix all padded tracks into single audio file using PyAV (same as Celery)."""
     ctx.log("mixdown_tracks: mixing padded tracks into single audio file")
 
-    paddings_result = ctx.task_output(process_paddings)
+    track_result = ctx.task_output(process_tracks)
     recording_result = ctx.task_output(get_recording)
-    padded_tracks = paddings_result.padded_tracks
+    padded_tracks = track_result.padded_tracks
 
     # Dynamic timeout: scales with track count and recording duration
     # Base 300s + 60s per track + 1s per 10s of recording
@@ -609,13 +561,27 @@ async def mixdown_tracks(input: PipelineInput, ctx: Context) -> MixdownResult:
 
     Path(output_path).unlink(missing_ok=True)
 
+    duration = duration_ms_callback_capture_container[0]
+
     async with fresh_db_connection():
-        from reflector.db.transcripts import transcripts_controller  # noqa: PLC0415
+        from reflector.db.transcripts import (  # noqa: PLC0415
+            TranscriptDuration,
+            transcripts_controller,
+        )
 
         transcript = await transcripts_controller.get_by_id(input.transcript_id)
         if transcript:
             await transcripts_controller.update(
-                transcript, {"audio_location": "storage"}
+                transcript, {"audio_location": "storage", "duration": duration}
+            )
+
+            duration_data = TranscriptDuration(duration=duration)
+            await append_event_and_broadcast(
+                input.transcript_id,
+                transcript,
+                "DURATION",
+                duration_data,
+                logger=logger,
             )
 
     ctx.log(f"mixdown_tracks complete: uploaded {file_size} bytes to {storage_path}")
@@ -696,7 +662,7 @@ async def generate_waveform(input: PipelineInput, ctx: Context) -> WaveformResul
 
 
 @daily_multitrack_pipeline.task(
-    parents=[process_transcriptions],
+    parents=[process_tracks],
     execution_timeout=timedelta(seconds=TIMEOUT_HEAVY),
     retries=3,
 )
@@ -705,8 +671,8 @@ async def detect_topics(input: PipelineInput, ctx: Context) -> TopicsResult:
     """Detect topics using parallel child workflows (one per chunk)."""
     ctx.log("detect_topics: analyzing transcript for topics")
 
-    transcriptions_result = ctx.task_output(process_transcriptions)
+    track_result = ctx.task_output(process_tracks)
-    words = transcriptions_result.all_words
+    words = track_result.all_words
 
     if not words:
         ctx.log("detect_topics: no words, returning empty topics")
@@ -1143,7 +1109,7 @@ async def identify_action_items(
 
 
 @daily_multitrack_pipeline.task(
-    parents=[generate_title, generate_recap, identify_action_items],
+    parents=[process_tracks, generate_title, generate_recap, identify_action_items],
     execution_timeout=timedelta(seconds=TIMEOUT_SHORT),
     retries=3,
 )
@@ -1156,15 +1122,10 @@ async def finalize(input: PipelineInput, ctx: Context) -> FinalizeResult:
     """
     ctx.log("finalize: saving transcript and setting status to 'ended'")
 
-    mixdown_result = ctx.task_output(mixdown_tracks)
+    track_result = ctx.task_output(process_tracks)
-    transcriptions_result = ctx.task_output(process_transcriptions)
-    paddings_result = ctx.task_output(process_paddings)
-
-    duration = mixdown_result.duration
-    all_words = transcriptions_result.all_words
 
     # Cleanup temporary padded S3 files (deferred until finalize for semantic parity with Celery)
-    created_padded_files = paddings_result.created_padded_files
+    created_padded_files = track_result.created_padded_files
     if created_padded_files:
         ctx.log(f"Cleaning up {len(created_padded_files)} temporary S3 files")
         storage = _spawn_storage()
@@ -1182,7 +1143,6 @@ async def finalize(input: PipelineInput, ctx: Context) -> FinalizeResult:
 
     async with fresh_db_connection():
         from reflector.db.transcripts import (  # noqa: PLC0415
-            TranscriptDuration,
             TranscriptText,
             transcripts_controller,
         )
@@ -1191,8 +1151,6 @@ async def finalize(input: PipelineInput, ctx: Context) -> FinalizeResult:
         if transcript is None:
             raise ValueError(f"Transcript {input.transcript_id} not found in database")
 
-        merged_transcript = TranscriptType(words=all_words, translation=None)
-
         await append_event_and_broadcast(
             input.transcript_id,
             transcript,
@@ -1204,21 +1162,15 @@ async def finalize(input: PipelineInput, ctx: Context) -> FinalizeResult:
             logger=logger,
         )
 
-        # Save duration and clear workflow_run_id (workflow completed successfully)
+        # Clear workflow_run_id (workflow completed successfully)
-        # Note: title/long_summary/short_summary already saved by their callbacks
+        # Note: title/long_summary/short_summary/duration already saved by their callbacks
         await transcripts_controller.update(
             transcript,
             {
-                "duration": duration,
                 "workflow_run_id": None,  # Clear on success - no need to resume
             },
         )
 
-        duration_data = TranscriptDuration(duration=duration)
-        await append_event_and_broadcast(
-            input.transcript_id, transcript, "DURATION", duration_data, logger=logger
-        )
-
         await set_status_and_broadcast(input.transcript_id, "ended", logger=logger)
 
         ctx.log(

server/reflector/hatchet/workflows/models.py

@@ -21,14 +21,12 @@ class ParticipantInfo(BaseModel):
 
 
 class PadTrackResult(BaseModel):
-    """Result from pad_track task.
+    """Result from pad_track task."""
 
-    If size=0, track required no padding and padded_key contains original S3 key.
+    padded_key: NonEmptyString  # S3 key (not presigned URL) - presign on demand to avoid stale URLs on replay
-    If size>0, track was padded and padded_key contains new padded file S3 key.
+    bucket_name: (
-    """
+        NonEmptyString | None
-
+    )  # None means use default transcript storage bucket
-    padded_key: NonEmptyString
-    bucket_name: NonEmptyString | None
     size: int
     track_index: int
 
@@ -61,25 +59,18 @@ class PaddedTrackInfo(BaseModel):
     """Info for a padded track - S3 key + bucket for on-demand presigning."""
 
     key: NonEmptyString
-    bucket_name: NonEmptyString | None
+    bucket_name: NonEmptyString | None  # None = use default storage bucket
-    track_index: int
 
 
-class ProcessPaddingsResult(BaseModel):
+class ProcessTracksResult(BaseModel):
-    """Result from process_paddings task."""
+    """Result from process_tracks task."""
-
-    padded_tracks: list[PaddedTrackInfo]
-    num_tracks: int
-    created_padded_files: list[NonEmptyString]
-
-
-class ProcessTranscriptionsResult(BaseModel):
-    """Result from process_transcriptions task."""
 
     all_words: list[Word]
+    padded_tracks: list[PaddedTrackInfo]  # S3 keys, not presigned URLs
     word_count: int
     num_tracks: int
     target_language: NonEmptyString
+    created_padded_files: list[NonEmptyString]
 
 
 class MixdownResult(BaseModel):

server/reflector/hatchet/workflows/padding_workflow.py

@@ -1,11 +1,9 @@
 """
 Hatchet child workflow: PaddingWorkflow
-Handles individual audio track padding only.
+Handles individual audio track padding via Modal.com backend.
 """
 
-import tempfile
 from datetime import timedelta
-from pathlib import Path
 
 import av
 from hatchet_sdk import Context
@@ -16,10 +14,7 @@ from reflector.hatchet.constants import TIMEOUT_AUDIO
 from reflector.hatchet.workflows.models import PadTrackResult
 from reflector.logger import logger
 from reflector.utils.audio_constants import PRESIGNED_URL_EXPIRATION_SECONDS
-from reflector.utils.audio_padding import (
+from reflector.utils.audio_padding import extract_stream_start_time_from_container
-    apply_audio_padding_to_file,
-    extract_stream_start_time_from_container,
-)
 
 
 class PaddingInput(BaseModel):
@@ -68,61 +63,83 @@ async def pad_track(input: PaddingInput, ctx: Context) -> PadTrackResult:
             bucket=input.bucket_name,
         )
 
+        # Extract start_time to determine if padding needed
         with av.open(source_url) as in_container:
-            with av.open(source_url) as in_container:
+            if in_container.duration:
-                if in_container.duration:
+                try:
-                    try:
+                    duration = timedelta(seconds=in_container.duration // 1_000_000)
-                        duration = timedelta(seconds=in_container.duration // 1_000_000)
+                    ctx.log(
-                        ctx.log(
+                        f"pad_track: track {input.track_index}, duration={duration}"
-                            f"pad_track: track {input.track_index}, duration={duration}"
+                    )
-                        )
+                except (ValueError, TypeError, OverflowError) as e:
-                    except (ValueError, TypeError, OverflowError) as e:
+                    ctx.log(
-                        ctx.log(
+                        f"pad_track: track {input.track_index}, duration error: {str(e)}"
-                            f"pad_track: track {input.track_index}, duration error: {str(e)}"
+                    )
-                        )
 
             start_time_seconds = extract_stream_start_time_from_container(
                 in_container, input.track_index, logger=logger
             )
 
-            if start_time_seconds <= 0:
+        if start_time_seconds <= 0:
-                logger.info(
+            logger.info(
-                    f"Track {input.track_index} requires no padding",
+                f"Track {input.track_index} requires no padding",
-                    track_index=input.track_index,
+                track_index=input.track_index,
-                )
+            )
-                return PadTrackResult(
+            return PadTrackResult(
-                    padded_key=input.s3_key,
+                padded_key=input.s3_key,
-                    bucket_name=input.bucket_name,
+                bucket_name=input.bucket_name,
-                    size=0,
+                size=0,
-                    track_index=input.track_index,
+                track_index=input.track_index,
-                )
+            )
 
-            storage_path = f"file_pipeline_hatchet/{input.transcript_id}/tracks/padded_{input.track_index}.webm"
+        storage_path = f"file_pipeline_hatchet/{input.transcript_id}/tracks/padded_{input.track_index}.webm"
 
-            with tempfile.NamedTemporaryFile(suffix=".webm", delete=False) as temp_file:
+        # Presign PUT URL for output (Modal will upload directly)
-                temp_path = temp_file.name
+        output_url = await storage.get_file_url(
+            storage_path,
+            operation="put_object",
+            expires_in=PRESIGNED_URL_EXPIRATION_SECONDS,
+        )
 
-            try:
+        import httpx  # noqa: PLC0415
-                apply_audio_padding_to_file(
-                    in_container,
-                    temp_path,
-                    start_time_seconds,
-                    input.track_index,
-                    logger=logger,
-                )
 
-                file_size = Path(temp_path).stat().st_size
+        from reflector.processors.audio_padding_modal import (  # noqa: PLC0415
+            AudioPaddingModalProcessor,
+        )
 
-                with open(temp_path, "rb") as padded_file:
+        try:
-                    await storage.put_file(storage_path, padded_file)
+            processor = AudioPaddingModalProcessor()
+            result = await processor.pad_track(
+                track_url=source_url,
+                output_url=output_url,
+                start_time_seconds=start_time_seconds,
+                track_index=input.track_index,
+            )
+            file_size = result.size
 
-                logger.info(
+            ctx.log(f"pad_track: Modal returned size={file_size}")
-                    f"Uploaded padded track to S3",
+        except httpx.HTTPStatusError as e:
-                    key=storage_path,
+            error_detail = e.response.text if hasattr(e.response, "text") else str(e)
-                    size=file_size,
+            logger.error(
-                )
+                "[Hatchet] Modal padding HTTP error",
-            finally:
+                transcript_id=input.transcript_id,
-                Path(temp_path).unlink(missing_ok=True)
+                track_index=input.track_index,
+                status_code=e.response.status_code if hasattr(e, "response") else None,
+                error=error_detail,
+                exc_info=True,
+            )
+            raise Exception(
+                f"Modal padding failed: HTTP {e.response.status_code}"
+            ) from e
+        except httpx.TimeoutException as e:
+            logger.error(
+                "[Hatchet] Modal padding timeout",
+                transcript_id=input.transcript_id,
+                track_index=input.track_index,
+                error=str(e),
+                exc_info=True,
+            )
+            raise Exception("Modal padding timeout") from e
 
         logger.info(
             "[Hatchet] pad_track complete",

server/reflector/hatchet/workflows/track_processing.py

@@ -0,0 +1,205 @@
+"""
+Hatchet child workflow: TrackProcessing
+
+Handles individual audio track processing: padding and transcription.
+Spawned dynamically by the main diarization pipeline for each track.
+
+Architecture note: This is a separate workflow (not inline tasks in DailyMultitrackPipeline)
+because Hatchet workflow DAGs are defined statically, but the number of tracks varies
+at runtime. Child workflow spawning via `aio_run()` + `asyncio.gather()` is the
+standard pattern for dynamic fan-out. See `process_tracks` in daily_multitrack_pipeline.py.
+
+Note: This file uses deferred imports (inside tasks) intentionally.
+Hatchet workers run in forked processes; fresh imports per task ensure
+storage/DB connections are not shared across forks.
+"""
+
+from datetime import timedelta
+
+import av
+from hatchet_sdk import Context
+from pydantic import BaseModel
+
+from reflector.hatchet.client import HatchetClientManager
+from reflector.hatchet.constants import TIMEOUT_AUDIO, TIMEOUT_HEAVY
+from reflector.hatchet.workflows.models import PadTrackResult, TranscribeTrackResult
+from reflector.logger import logger
+from reflector.utils.audio_constants import PRESIGNED_URL_EXPIRATION_SECONDS
+from reflector.utils.audio_padding import extract_stream_start_time_from_container
+
+
+class TrackInput(BaseModel):
+    """Input for individual track processing."""
+
+    track_index: int
+    s3_key: str
+    bucket_name: str
+    transcript_id: str
+    language: str = "en"
+
+
+hatchet = HatchetClientManager.get_client()
+
+track_workflow = hatchet.workflow(name="TrackProcessing", input_validator=TrackInput)
+
+
+@track_workflow.task(execution_timeout=timedelta(seconds=TIMEOUT_AUDIO), retries=3)
+async def pad_track(input: TrackInput, ctx: Context) -> PadTrackResult:
+    """Pad single audio track with silence for alignment.
+
+    Extracts stream.start_time from WebM container metadata and applies
+    silence padding using PyAV filter graph (adelay).
+    """
+    ctx.log(f"pad_track: track {input.track_index}, s3_key={input.s3_key}")
+    logger.info(
+        "[Hatchet] pad_track",
+        track_index=input.track_index,
+        s3_key=input.s3_key,
+        transcript_id=input.transcript_id,
+    )
+
+    try:
+        # Create fresh storage instance to avoid aioboto3 fork issues
+        from reflector.settings import settings  # noqa: PLC0415
+        from reflector.storage.storage_aws import AwsStorage  # noqa: PLC0415
+
+        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,
+        )
+
+        source_url = await storage.get_file_url(
+            input.s3_key,
+            operation="get_object",
+            expires_in=PRESIGNED_URL_EXPIRATION_SECONDS,
+            bucket=input.bucket_name,
+        )
+
+        with av.open(source_url) as in_container:
+            start_time_seconds = extract_stream_start_time_from_container(
+                in_container, input.track_index, logger=logger
+            )
+
+        # If no padding needed, return original S3 key
+        if start_time_seconds <= 0:
+            logger.info(
+                f"Track {input.track_index} requires no padding",
+                track_index=input.track_index,
+            )
+            return PadTrackResult(
+                padded_key=input.s3_key,
+                bucket_name=input.bucket_name,
+                size=0,
+                track_index=input.track_index,
+            )
+
+        storage_path = f"file_pipeline_hatchet/{input.transcript_id}/tracks/padded_{input.track_index}.webm"
+
+        # Presign PUT URL for output (Modal uploads directly)
+        output_url = await storage.get_file_url(
+            storage_path,
+            operation="put_object",
+            expires_in=PRESIGNED_URL_EXPIRATION_SECONDS,
+        )
+
+        from reflector.processors.audio_padding_modal import (  # noqa: PLC0415
+            AudioPaddingModalProcessor,
+        )
+
+        processor = AudioPaddingModalProcessor()
+        result = await processor.pad_track(
+            track_url=source_url,
+            output_url=output_url,
+            start_time_seconds=start_time_seconds,
+            track_index=input.track_index,
+        )
+        file_size = result.size
+
+        ctx.log(f"pad_track complete: track {input.track_index} -> {storage_path}")
+        logger.info(
+            "[Hatchet] pad_track complete",
+            track_index=input.track_index,
+            padded_key=storage_path,
+        )
+
+        # Return S3 key (not presigned URL) - consumer tasks presign on demand
+        # This avoids stale URLs when workflow is replayed
+        return PadTrackResult(
+            padded_key=storage_path,
+            bucket_name=None,  # None = use default transcript storage bucket
+            size=file_size,
+            track_index=input.track_index,
+        )
+
+    except Exception as e:
+        logger.error("[Hatchet] pad_track failed", error=str(e), exc_info=True)
+        raise
+
+
+@track_workflow.task(
+    parents=[pad_track], execution_timeout=timedelta(seconds=TIMEOUT_HEAVY), retries=3
+)
+async def transcribe_track(input: TrackInput, ctx: Context) -> TranscribeTrackResult:
+    """Transcribe audio track using GPU (Modal.com) or local Whisper."""
+    ctx.log(f"transcribe_track: track {input.track_index}, language={input.language}")
+    logger.info(
+        "[Hatchet] transcribe_track",
+        track_index=input.track_index,
+        language=input.language,
+    )
+
+    try:
+        pad_result = ctx.task_output(pad_track)
+        padded_key = pad_result.padded_key
+        bucket_name = pad_result.bucket_name
+
+        if not padded_key:
+            raise ValueError("Missing padded_key from pad_track")
+
+        # Presign URL on demand (avoids stale URLs on workflow replay)
+        from reflector.settings import settings  # noqa: PLC0415
+        from reflector.storage.storage_aws import AwsStorage  # noqa: PLC0415
+
+        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,
+        )
+
+        audio_url = await storage.get_file_url(
+            padded_key,
+            operation="get_object",
+            expires_in=PRESIGNED_URL_EXPIRATION_SECONDS,
+            bucket=bucket_name,
+        )
+
+        from reflector.pipelines.transcription_helpers import (  # noqa: PLC0415
+            transcribe_file_with_processor,
+        )
+
+        transcript = await transcribe_file_with_processor(audio_url, input.language)
+
+        # Tag all words with speaker index
+        for word in transcript.words:
+            word.speaker = input.track_index
+
+        ctx.log(
+            f"transcribe_track complete: track {input.track_index}, {len(transcript.words)} words"
+        )
+        logger.info(
+            "[Hatchet] transcribe_track complete",
+            track_index=input.track_index,
+            word_count=len(transcript.words),
+        )
+
+        return TranscribeTrackResult(
+            words=transcript.words,
+            track_index=input.track_index,
+        )
+
+    except Exception as e:
+        logger.error("[Hatchet] transcribe_track failed", error=str(e), exc_info=True)
+        raise

server/reflector/hatchet/workflows/transcription_workflow.py

@@ -1,98 +0,0 @@
-"""
-Hatchet child workflow: TranscriptionWorkflow
-Handles individual audio track transcription only.
-"""
-
-from datetime import timedelta
-
-from hatchet_sdk import Context
-from pydantic import BaseModel
-
-from reflector.hatchet.client import HatchetClientManager
-from reflector.hatchet.constants import TIMEOUT_HEAVY
-from reflector.hatchet.workflows.models import TranscribeTrackResult
-from reflector.logger import logger
-from reflector.utils.audio_constants import PRESIGNED_URL_EXPIRATION_SECONDS
-
-
-class TranscriptionInput(BaseModel):
-    """Input for individual track transcription."""
-
-    track_index: int
-    padded_key: str  # S3 key from padding step
-    bucket_name: str | None  # None = use default bucket
-    language: str = "en"
-
-
-hatchet = HatchetClientManager.get_client()
-
-transcription_workflow = hatchet.workflow(
-    name="TranscriptionWorkflow", input_validator=TranscriptionInput
-)
-
-
-@transcription_workflow.task(
-    execution_timeout=timedelta(seconds=TIMEOUT_HEAVY), retries=3
-)
-async def transcribe_track(
-    input: TranscriptionInput, ctx: Context
-) -> TranscribeTrackResult:
-    """Transcribe audio track using GPU (Modal.com) or local Whisper."""
-    ctx.log(f"transcribe_track: track {input.track_index}, language={input.language}")
-    logger.info(
-        "[Hatchet] transcribe_track",
-        track_index=input.track_index,
-        language=input.language,
-    )
-
-    try:
-        from reflector.settings import settings  # noqa: PLC0415
-        from reflector.storage.storage_aws import AwsStorage  # noqa: PLC0415
-
-        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,
-        )
-
-        audio_url = await storage.get_file_url(
-            input.padded_key,
-            operation="get_object",
-            expires_in=PRESIGNED_URL_EXPIRATION_SECONDS,
-            bucket=input.bucket_name,
-        )
-
-        from reflector.pipelines.transcription_helpers import (  # noqa: PLC0415
-            transcribe_file_with_processor,
-        )
-
-        transcript = await transcribe_file_with_processor(audio_url, input.language)
-
-        for word in transcript.words:
-            word.speaker = input.track_index
-
-        ctx.log(
-            f"transcribe_track complete: track {input.track_index}, {len(transcript.words)} words"
-        )
-        logger.info(
-            "[Hatchet] transcribe_track complete",
-            track_index=input.track_index,
-            word_count=len(transcript.words),
-        )
-
-        return TranscribeTrackResult(
-            words=transcript.words,
-            track_index=input.track_index,
-        )
-
-    except Exception as e:
-        logger.error(
-            "[Hatchet] transcribe_track failed",
-            track_index=input.track_index,
-            padded_key=input.padded_key,
-            language=input.language,
-            error=str(e),
-            exc_info=True,
-        )
-        raise

server/reflector/processors/audio_padding_modal.py

@@ -0,0 +1,113 @@
+"""
+Modal.com backend for audio padding.
+"""
+
+import asyncio
+import os
+
+import httpx
+from pydantic import BaseModel
+
+from reflector.hatchet.constants import TIMEOUT_AUDIO
+from reflector.logger import logger
+
+
+class PaddingResponse(BaseModel):
+    size: int
+    cancelled: bool = False
+
+
+class AudioPaddingModalProcessor:
+    """Audio padding processor using Modal.com CPU backend via HTTP."""
+
+    def __init__(
+        self, padding_url: str | None = None, modal_api_key: str | None = None
+    ):
+        self.padding_url = padding_url or os.getenv("PADDING_URL")
+        if not self.padding_url:
+            raise ValueError(
+                "PADDING_URL required to use AudioPaddingModalProcessor. "
+                "Set PADDING_URL environment variable or pass padding_url parameter."
+            )
+
+        self.modal_api_key = modal_api_key or os.getenv("MODAL_API_KEY")
+
+    async def pad_track(
+        self,
+        track_url: str,
+        output_url: str,
+        start_time_seconds: float,
+        track_index: int,
+    ) -> PaddingResponse:
+        """Pad audio track with silence via Modal backend.
+
+        Args:
+            track_url: Presigned GET URL for source audio track
+            output_url: Presigned PUT URL for output WebM
+            start_time_seconds: Amount of silence to prepend
+            track_index: Track index for logging
+        """
+        if not track_url:
+            raise ValueError("track_url cannot be empty")
+        if start_time_seconds <= 0:
+            raise ValueError(
+                f"start_time_seconds must be positive, got {start_time_seconds}"
+            )
+
+        log = logger.bind(track_index=track_index, padding_seconds=start_time_seconds)
+        log.info("Sending Modal padding HTTP request")
+
+        url = f"{self.padding_url}/pad"
+
+        headers = {}
+        if self.modal_api_key:
+            headers["Authorization"] = f"Bearer {self.modal_api_key}"
+
+        try:
+            async with httpx.AsyncClient(timeout=TIMEOUT_AUDIO) as client:
+                response = await client.post(
+                    url,
+                    headers=headers,
+                    json={
+                        "track_url": track_url,
+                        "output_url": output_url,
+                        "start_time_seconds": start_time_seconds,
+                        "track_index": track_index,
+                    },
+                    follow_redirects=True,
+                )
+
+                if response.status_code != 200:
+                    error_body = response.text
+                    log.error(
+                        "Modal padding API error",
+                        status_code=response.status_code,
+                        error_body=error_body,
+                    )
+
+                response.raise_for_status()
+                result = response.json()
+
+            # Check if work was cancelled
+            if result.get("cancelled"):
+                log.warning("Modal padding was cancelled by disconnect detection")
+                raise asyncio.CancelledError(
+                    "Padding cancelled due to client disconnect"
+                )
+
+            log.info("Modal padding complete", size=result["size"])
+            return PaddingResponse(**result)
+        except asyncio.CancelledError:
+            log.warning(
+                "Modal padding cancelled (Hatchet timeout, disconnect detected on Modal side)"
+            )
+            raise
+        except httpx.TimeoutException as e:
+            log.error("Modal padding timeout", error=str(e), exc_info=True)
+            raise Exception(f"Modal padding timeout: {e}") from e
+        except httpx.HTTPStatusError as e:
+            log.error("Modal padding HTTP error", error=str(e), exc_info=True)
+            raise Exception(f"Modal padding HTTP error: {e}") from e
+        except Exception as e:
+            log.error("Modal padding unexpected error", error=str(e), exc_info=True)
+            raise

server/reflector/settings.py

@@ -98,6 +98,10 @@ class Settings(BaseSettings):
     # 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
+
     # Sentry
     SENTRY_DSN: str | None = None
 

server/reflector/utils/audio_constants.py

@@ -5,7 +5,9 @@ Used by both Hatchet workflows and Celery pipelines for consistent audio encodin
 """
 
 # Opus codec settings
+# ref B0F71CE8-FC59-4AA5-8414-DAFB836DB711
 OPUS_STANDARD_SAMPLE_RATE = 48000
+# ref B0F71CE8-FC59-4AA5-8414-DAFB836DB711
 OPUS_DEFAULT_BIT_RATE = 128000  # 128kbps for good speech quality
 
 # S3 presigned URL expiration

server/uv.lock

@@ -159,21 +159,20 @@ wheels = [
 
 [[package]]
 name = "aiortc"
-version = "1.13.0"
+version = "1.14.0"
 source = { registry = "https://pypi.org/simple" }
 dependencies = [
     { name = "aioice" },
     { name = "av" },
-    { name = "cffi" },
     { name = "cryptography" },
     { name = "google-crc32c" },
     { name = "pyee" },
     { name = "pylibsrtp" },
     { name = "pyopenssl" },
 ]
-sdist = { url = "https://files.pythonhosted.org/packages/62/03/bc947d74c548e0c17cf94e5d5bdacaed0ee9e5b2bb7b8b8cf1ac7a7c01ec/aiortc-1.13.0.tar.gz", hash = "sha256:5d209975c22d0910fb5a0f0e2caa828f2da966c53580f7c7170ac3a16a871620", size = 1179894 }
+sdist = { url = "https://files.pythonhosted.org/packages/51/9c/4e027bfe0195de0442da301e2389329496745d40ae44d2d7c4571c4290ce/aiortc-1.14.0.tar.gz", hash = "sha256:adc8a67ace10a085721e588e06a00358ed8eaf5f6b62f0a95358ff45628dd762", size = 1180864 }
 wheels = [
-    { url = "https://files.pythonhosted.org/packages/87/29/765633cab5f1888890f5f172d1d53009b9b14e079cdfa01a62d9896a9ea9/aiortc-1.13.0-py3-none-any.whl", hash = "sha256:9ccccec98796f6a96bd1c3dd437a06da7e0f57521c96bd56e4b965a91b03a0a0", size = 92910 },
+    { url = "https://files.pythonhosted.org/packages/57/ab/31646a49209568cde3b97eeade0d28bb78b400e6645c56422c101df68932/aiortc-1.14.0-py3-none-any.whl", hash = "sha256:4b244d7e482f4e1f67e685b3468269628eca1ec91fa5b329ab517738cfca086e", size = 93183 },
 ]
 
 [[package]]
@@ -327,28 +326,24 @@ wheels = [
 
 [[package]]
 name = "av"
-version = "14.4.0"
+version = "16.1.0"
 source = { registry = "https://pypi.org/simple" }
-sdist = { url = "https://files.pythonhosted.org/packages/86/f6/0b473dab52dfdea05f28f3578b1c56b6c796ce85e76951bab7c4e38d5a74/av-14.4.0.tar.gz", hash = "sha256:3ecbf803a7fdf67229c0edada0830d6bfaea4d10bfb24f0c3f4e607cd1064b42", size = 3892203 }
+sdist = { url = "https://files.pythonhosted.org/packages/78/cd/3a83ffbc3cc25b39721d174487fb0d51a76582f4a1703f98e46170ce83d4/av-16.1.0.tar.gz", hash = "sha256:a094b4fd87a3721dacf02794d3d2c82b8d712c85b9534437e82a8a978c175ffd", size = 4285203 }
 wheels = [
-    { url = "https://files.pythonhosted.org/packages/18/8a/d57418b686ffd05fabd5a0a9cfa97e63b38c35d7101af00e87c51c8cc43c/av-14.4.0-cp311-cp311-macosx_12_0_arm64.whl", hash = "sha256:5b21d5586a88b9fce0ab78e26bd1c38f8642f8e2aad5b35e619f4d202217c701", size = 19965048 },
+    { url = "https://files.pythonhosted.org/packages/48/d0/b71b65d1b36520dcb8291a2307d98b7fc12329a45614a303ff92ada4d723/av-16.1.0-cp311-cp311-macosx_11_0_x86_64.whl", hash = "sha256:e88ad64ee9d2b9c4c5d891f16c22ae78e725188b8926eb88187538d9dd0b232f", size = 26927747 },
-    { url = "https://files.pythonhosted.org/packages/f5/aa/3f878b0301efe587e9b07bb773dd6b47ef44ca09a3cffb4af50c08a170f3/av-14.4.0-cp311-cp311-macosx_12_0_x86_64.whl", hash = "sha256:cf8762d90b0f94a20c9f6e25a94f1757db5a256707964dfd0b1d4403e7a16835", size = 23750064 },
+    { url = "https://files.pythonhosted.org/packages/2f/79/720a5a6ccdee06eafa211b945b0a450e3a0b8fc3d12922f0f3c454d870d2/av-16.1.0-cp311-cp311-macosx_14_0_arm64.whl", hash = "sha256:cb296073fa6935724de72593800ba86ae49ed48af03960a4aee34f8a611f442b", size = 21492232 },
-    { url = "https://files.pythonhosted.org/packages/9a/b4/6fe94a31f9ed3a927daa72df67c7151968587106f30f9f8fcd792b186633/av-14.4.0-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:c0ac9f08920c7bbe0795319689d901e27cb3d7870b9a0acae3f26fc9daa801a6", size = 33648775 },
+    { url = "https://files.pythonhosted.org/packages/8e/4f/a1ba8d922f2f6d1a3d52419463ef26dd6c4d43ee364164a71b424b5ae204/av-16.1.0-cp311-cp311-manylinux_2_28_aarch64.whl", hash = "sha256:720edd4d25aa73723c1532bb0597806d7b9af5ee34fc02358782c358cfe2f879", size = 39291737 },
-    { url = "https://files.pythonhosted.org/packages/6c/f3/7f3130753521d779450c935aec3f4beefc8d4645471159f27b54e896470c/av-14.4.0-cp311-cp311-manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:a56d9ad2afdb638ec0404e962dc570960aae7e08ae331ad7ff70fbe99a6cf40e", size = 32216915 },
+    { url = "https://files.pythonhosted.org/packages/1a/31/fc62b9fe8738d2693e18d99f040b219e26e8df894c10d065f27c6b4f07e3/av-16.1.0-cp311-cp311-manylinux_2_28_x86_64.whl", hash = "sha256:c7f2bc703d0df260a1fdf4de4253c7f5500ca9fc57772ea241b0cb241bcf972e", size = 40846822 },
-    { url = "https://files.pythonhosted.org/packages/f8/9a/8ffabfcafb42154b4b3a67d63f9b69e68fa8c34cb39ddd5cb813dd049ed4/av-14.4.0-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:6bed513cbcb3437d0ae47743edc1f5b4a113c0b66cdd4e1aafc533abf5b2fbf2", size = 35287279 },
+    { url = "https://files.pythonhosted.org/packages/53/10/ab446583dbce730000e8e6beec6ec3c2753e628c7f78f334a35cad0317f4/av-16.1.0-cp311-cp311-musllinux_1_2_aarch64.whl", hash = "sha256:d69c393809babada7d54964d56099e4b30a3e1f8b5736ca5e27bd7be0e0f3c83", size = 40675604 },
-    { url = "https://files.pythonhosted.org/packages/ad/11/7023ba0a2ca94a57aedf3114ab8cfcecb0819b50c30982a4c5be4d31df41/av-14.4.0-cp311-cp311-musllinux_1_2_aarch64.whl", hash = "sha256:d030c2d3647931e53d51f2f6e0fcf465263e7acf9ec6e4faa8dbfc77975318c3", size = 36294683 },
+    { url = "https://files.pythonhosted.org/packages/31/d7/1003be685277005f6d63fd9e64904ee222fe1f7a0ea70af313468bb597db/av-16.1.0-cp311-cp311-musllinux_1_2_x86_64.whl", hash = "sha256:441892be28582356d53f282873c5a951592daaf71642c7f20165e3ddcb0b4c63", size = 42015955 },
-    { url = "https://files.pythonhosted.org/packages/3d/fa/b8ac9636bd5034e2b899354468bef9f4dadb067420a16d8a493a514b7817/av-14.4.0-cp311-cp311-musllinux_1_2_i686.whl", hash = "sha256:1cc21582a4f606271d8c2036ec7a6247df0831050306c55cf8a905701d0f0474", size = 34552391 },
+    { url = "https://files.pythonhosted.org/packages/2f/4a/fa2a38ee9306bf4579f556f94ecbc757520652eb91294d2a99c7cf7623b9/av-16.1.0-cp311-cp311-win_amd64.whl", hash = "sha256:273a3e32de64819e4a1cd96341824299fe06f70c46f2288b5dc4173944f0fd62", size = 31750339 },
-    { url = "https://files.pythonhosted.org/packages/fb/29/0db48079c207d1cba7a2783896db5aec3816e17de55942262c244dffbc0f/av-14.4.0-cp311-cp311-musllinux_1_2_x86_64.whl", hash = "sha256:ce7c9cd452153d36f1b1478f904ed5f9ab191d76db873bdd3a597193290805d4", size = 37265250 },
+    { url = "https://files.pythonhosted.org/packages/9c/84/2535f55edcd426cebec02eb37b811b1b0c163f26b8d3f53b059e2ec32665/av-16.1.0-cp312-cp312-macosx_11_0_x86_64.whl", hash = "sha256:640f57b93f927fba8689f6966c956737ee95388a91bd0b8c8b5e0481f73513d6", size = 26945785 },
-    { url = "https://files.pythonhosted.org/packages/1c/55/715858c3feb7efa4d667ce83a829c8e6ee3862e297fb2b568da3f968639d/av-14.4.0-cp311-cp311-win_amd64.whl", hash = "sha256:fd261e31cc6b43ca722f80656c39934199d8f2eb391e0147e704b6226acebc29", size = 27925845 },
+    { url = "https://files.pythonhosted.org/packages/b6/17/ffb940c9e490bf42e86db4db1ff426ee1559cd355a69609ec1efe4d3a9eb/av-16.1.0-cp312-cp312-macosx_14_0_arm64.whl", hash = "sha256:ae3fb658eec00852ebd7412fdc141f17f3ddce8afee2d2e1cf366263ad2a3b35", size = 21481147 },
-    { url = "https://files.pythonhosted.org/packages/a6/75/b8641653780336c90ba89e5352cac0afa6256a86a150c7703c0b38851c6d/av-14.4.0-cp312-cp312-macosx_12_0_arm64.whl", hash = "sha256:a53e682b239dd23b4e3bc9568cfb1168fc629ab01925fdb2e7556eb426339e94", size = 19954125 },
+    { url = "https://files.pythonhosted.org/packages/15/c1/e0d58003d2d83c3921887d5c8c9b8f5f7de9b58dc2194356a2656a45cfdc/av-16.1.0-cp312-cp312-manylinux_2_28_aarch64.whl", hash = "sha256:27ee558d9c02a142eebcbe55578a6d817fedfde42ff5676275504e16d07a7f86", size = 39517197 },
-    { url = "https://files.pythonhosted.org/packages/99/e6/37fe6fa5853a48d54d749526365780a63a4bc530be6abf2115e3a21e292a/av-14.4.0-cp312-cp312-macosx_12_0_x86_64.whl", hash = "sha256:5aa0b901751a32703fa938d2155d56ce3faf3630e4a48d238b35d2f7e49e5395", size = 23751479 },
+    { url = "https://files.pythonhosted.org/packages/32/77/787797b43475d1b90626af76f80bfb0c12cfec5e11eafcfc4151b8c80218/av-16.1.0-cp312-cp312-manylinux_2_28_x86_64.whl", hash = "sha256:7ae547f6d5fa31763f73900d43901e8c5fa6367bb9a9840978d57b5a7ae14ed2", size = 41174337 },
-    { url = "https://files.pythonhosted.org/packages/f7/75/9a5f0e6bda5f513b62bafd1cff2b495441a8b07ab7fb7b8e62f0c0d1683f/av-14.4.0-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:a3b316fed3597675fe2aacfed34e25fc9d5bb0196dc8c0b014ae5ed4adda48de", size = 33801401 },
+    { url = "https://files.pythonhosted.org/packages/8e/ac/d90df7f1e3b97fc5554cf45076df5045f1e0a6adf13899e10121229b826c/av-16.1.0-cp312-cp312-musllinux_1_2_aarch64.whl", hash = "sha256:8cf065f9d438e1921dc31fc7aa045790b58aee71736897866420d80b5450f62a", size = 40817720 },
-    { url = "https://files.pythonhosted.org/packages/6a/c9/e4df32a2ad1cb7f3a112d0ed610c5e43c89da80b63c60d60e3dc23793ec0/av-14.4.0-cp312-cp312-manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:a587b5c5014c3c0e16143a0f8d99874e46b5d0c50db6111aa0b54206b5687c81", size = 32364330 },
+    { url = "https://files.pythonhosted.org/packages/80/6f/13c3a35f9dbcebafd03fe0c4cbd075d71ac8968ec849a3cfce406c35a9d2/av-16.1.0-cp312-cp312-musllinux_1_2_x86_64.whl", hash = "sha256:a345877a9d3cc0f08e2bc4ec163ee83176864b92587afb9d08dff50f37a9a829", size = 42267396 },
-    { url = "https://files.pythonhosted.org/packages/ca/f0/64e7444a41817fde49a07d0239c033f7e9280bec4a4bb4784f5c79af95e6/av-14.4.0-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:10d53f75e8ac1ec8877a551c0db32a83c0aaeae719d05285281eaaba211bbc30", size = 35519508 },
+    { url = "https://files.pythonhosted.org/packages/c8/b9/275df9607f7fb44317ccb1d4be74827185c0d410f52b6e2cd770fe209118/av-16.1.0-cp312-cp312-win_amd64.whl", hash = "sha256:f49243b1d27c91cd8c66fdba90a674e344eb8eb917264f36117bf2b6879118fd", size = 31752045 },
-    { url = "https://files.pythonhosted.org/packages/c2/a8/a370099daa9033a3b6f9b9bd815304b3d8396907a14d09845f27467ba138/av-14.4.0-cp312-cp312-musllinux_1_2_aarch64.whl", hash = "sha256:c8558cfde79dd8fc92d97c70e0f0fa8c94c7a66f68ae73afdf58598f0fe5e10d", size = 36448593 },
-    { url = "https://files.pythonhosted.org/packages/27/bb/edb6ceff8fa7259cb6330c51dbfbc98dd1912bd6eb5f7bc05a4bb14a9d6e/av-14.4.0-cp312-cp312-musllinux_1_2_i686.whl", hash = "sha256:455b6410dea0ab2d30234ffb28df7d62ca3cdf10708528e247bec3a4cdcced09", size = 34701485 },
-    { url = "https://files.pythonhosted.org/packages/a7/8a/957da1f581aa1faa9a5dfa8b47ca955edb47f2b76b949950933b457bfa1d/av-14.4.0-cp312-cp312-musllinux_1_2_x86_64.whl", hash = "sha256:1661efbe9d975f927b8512d654704223d936f39016fad2ddab00aee7c40f412c", size = 37521981 },
-    { url = "https://files.pythonhosted.org/packages/28/76/3f1cf0568592f100fd68eb40ed8c491ce95ca3c1378cc2d4c1f6d1bd295d/av-14.4.0-cp312-cp312-win_amd64.whl", hash = "sha256:fbbeef1f421a3461086853d6464ad5526b56ffe8ccb0ab3fd0a1f121dfbf26ad", size = 27925944 },
 ]
 
 [[package]]
@@ -3267,7 +3262,7 @@ requires-dist = [
     { name = "aiohttp-cors", specifier = ">=0.7.0" },
     { name = "aiortc", specifier = ">=1.5.0" },
     { name = "alembic", specifier = ">=1.11.3" },
-    { name = "av", specifier = ">=10.0.0" },
+    { name = "av", specifier = ">=15.0.0" },
     { name = "celery", specifier = ">=5.3.4" },
     { name = "databases", extras = ["aiosqlite", "asyncpg"], specifier = ">=0.7.0" },
     { name = "fastapi", extras = ["standard"], specifier = ">=0.100.1" },