Without charset specification, browsers/tools interpret text/vtt as
Latin-1, causing UTF-8 multi-byte characters like ♪ (U+266A) to render
as garbled text (♪).
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
SDH (sdh_captions_vtt) was missing from the Downloads page type labels
and filename extensions map.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Set staleTime=0 on useAccessibleVideoEditState so edit-state is
immediately refetched when invalidated after render completes
- Force prevJobStatusRef='rendering_qc' at render start so the
pending_qc completion effect always fires, even when fast renders
complete before the 10s polling interval catches rendering_qc status
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Rewrote VTT translation to two-step (text-only → Gemini → apply to original timestamps) preventing caption timing desync
- Added polling fallback for all processing states and Safari visibilitychange WebSocket reconnect
- Added 11 new TTS languages (cs, da, fi, hu, no, sk, sv, es-419, pt-BR, fr-CA)
- Updated caption/AD prompts to DCMP Captioning Key & Description Key standards (line splitting, ♪ music notation, italic tags, caption positioning, ethics guidelines)
- Added descriptive transcript generation (WCAG 2.1 §1.2.1) combining captions + AD into plain text
- Fixed amix normalize=0 to prevent audio loss in rendered videos
- Fixed AD re-timing double-count when source_ms is None
- Fixed cue block numbering to be 1-based in VttEditor and Timeline Preview
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
**BUG-1 & BUG-2 — Wrong audio plays after re-render / MP3 doesn't match text**
Root cause: audio files were named by index (cue_0.mp3, cue_1.mp3). When a cue
was inserted or deleted, all following indices shifted but old MP3 files kept
their original names, so re-render would play the wrong audio for the wrong cue.
Fix: renamed files to cue_N_CONTENTHASH.mp3 and introduced an ad_cue_manifest
stored in the job document that maps each cue index to its correct GCS URI.
Re-render now reads from the manifest instead of guessing by filename.
Also: editing AD cue text in the VTT editor now automatically queues TTS
regeneration for changed cues — no more silent mismatches.
**BUG-3 — App crash / state desync when uploading VTT or clearing TTS queue**
Fixed handleVttFileUpload to only update local editor state after the server
confirms the save — previously local state was updated first, so a network
error left the UI showing content that wasn't actually saved.
Fixed handleClearRegenerationQueue to only remove items from local state if
the server removal succeeded — previously all items were cleared regardless.
**BUG-4 — AI generates different audio descriptions every time**
Added GenerateContentConfig(temperature=0.2, top_p=0.8, top_k=40) to the
Gemini API call so output is more consistent across runs.
**BUG-5 — On-screen text inconsistently described**
Strengthened the AI prompt rule from a vague suggestion to a mandatory
requirement with an explicit format: "Text on screen reads: [exact text]".
Applied to both gemini_ingestion.md and gemini_ingestion_targeted.md.
**BUG-6 — No notification when re-render finishes**
Added rendering_qc toast notification and a dismissible green banner that
appears in QCDetail when re-render transitions to pending_qc. The banner
auto-dismisses after 10 seconds. Also increased WebSocket reconnect attempts
from 5 to 15 and capped backoff at 60s to prevent falling back to manual refresh.
**BUG-7 — Timeline preview looks accurate but isn't after edits**
Added isStale prop to TimelinePreview. The timeline now shows an amber tint
and "Preview may be outdated" label whenever there are unsaved pause point
changes, pending TTS regenerations, or a new VTT has been uploaded.
**BUG-8 — ElevenLabs API errors break TTS with no fallback**
Added try/except fallback chain in _synthesize_single_cue: if the configured
provider fails, it automatically retries with google, then gemini.
**BUG-9 — Concurrent re-render requests cause race conditions**
Made the PENDING_QC → RENDERING_QC status transition conditional (only
succeeds if the job is still in PENDING_QC). Returns HTTP 409 if a re-render
is already in progress. The completion transition back to PENDING_QC is also
conditional so a cancelled/overridden render doesn't corrupt job state.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Microsoft can return different email casings for the same user (e.g.
VadymSamoilenko@... vs vadymsamoilenko@...). The previous case-sensitive
find_one would miss the existing user, then fail on insert_one with a
duplicate key error on the _id field (ms-{sub[:20]}).
Fix: look up by _id first (deterministic from Microsoft sub), then fall
back to case-insensitive email regex for local-to-Microsoft migrations.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
SDH captions extend standard VTT with speaker identification labels,
sound effects [PHONE RINGS], music notation ♪, and off-screen indicators.
- Add sdh_vtt flag to RequestedOutputs model and frontend form
- Add sdh_captions_vtt_gcs field to LangOutput model
- Inject SDH generation instructions into both Gemini prompts via
{SDH_FIELD} and {SDH_GUIDELINES} placeholders when requested
- Upload sdh_captions.vtt to GCS in ingest task
- Pass SDH through video_native translation (Gemini generates it directly)
and traditional translation (translate source SDH VTT via Gemini)
- Expose sdh_captions_vtt in downloads endpoint and bulk zip export
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Add brand_context field (job model, API, frontend form) so clients can
list brand names present in their video; Gemini uses these names instead
of generic descriptors (e.g. "Sellotape" not "sticky tape")
- Add ethical guidelines section to both Gemini prompts covering
person-first language, consistent race/gender description only when
plot-relevant, no guessing at unconfirmed identity
- Revamp audio description rules: priority ordering (essential →
high-priority → time-permitting), pre-teaching placement, no cinematic
jargon, succinct style replacing the former "20% longer" instruction
- Thread brand_context through full stack: routes → job doc → ingest
task → translate task → both Gemini prompt templates
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Old pause_points in edit_state always overrode new VTT cue timings
during re-render, making AD VTT upload for timing adjustments
non-functional. Clear pause_points and video_segments on AD VTT
upload so re-render falls back to the new cue start times.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Fixes race condition where timeline preview never updated after AD VTT
re-render: useJob now polls every 5s during rendering_qc, and QCDetail
invalidates the edit-state query when the job transitions back to pending_qc.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
adVttUploaded was missing from the hasChanges condition in
RerenderControls, leaving the button greyed out after an upload.
Pass the flag as a prop and include it in hasChanges; also show
an "Audio Description script was replaced" status message.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Reviewers can now download individual job assets (source video,
captions VTT, AD VTT, AD MP3, accessible video) directly from the
QC detail page. They can also replace captions or AD VTT scripts by
uploading a revised file, with a banner prompting re-render when the
AD script is replaced.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
useJobDownloads now accepts jobStatus and disables the query when the
job is in created/ingesting/ai_processing, preventing spurious 400s
from /jobs/{id}/downloads before any outputs exist.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- elevenlabs_voices.py: re-raise exception on first fetch failure
(empty cache) instead of silently returning empty list
- routes_tts.py: catch get_voices() exception and return available=False
with the error detail; add optional error field to ProviderVoicesResponse
- VoiceSelector: show actual API error message when available=false
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Extract actual error message from blob response in previewVoice so
users see the real API error instead of generic "Failed to generate preview"
- VoicePreviewButton now reads err.message from thrown Error objects
- Add available: bool field to ProviderVoicesResponse; returns false
when ELEVENLABS_API_KEY is not configured so the frontend can react
proactively instead of hitting a 400 on preview
- VoiceSelector shows a descriptive config warning when available=false
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Add dynamic ElevenLabs voice catalog with provider toggle in the UI,
allowing users to browse ElevenLabs voices, configure stability and
similarity boost settings, and preview/synthesize with ElevenLabs TTS.
Backend:
- New elevenlabs_voices.py service with 1-hour cached API fetching
- TTS routes support ?provider= query param for voices and options
- Preview endpoint routes to ElevenLabs or Gemini based on provider
- stability/similarity_boost params flow through TTS synthesis pipeline
- TTSPreferences model extended with ElevenLabs-specific fields
- Deprecated hardcoded elevenlabs_voices config (now fetched dynamically)
Frontend:
- Provider toggle (Gemini/ElevenLabs) in VoiceSelector
- ElevenLabsSettingsPanel with stability and similarity boost sliders
- VoicePreviewButton supports provider-specific preview parameters
- API client passes provider param to voices, options, and preview endpoints
- New VoiceInfo, ProviderVoicesResponse, ProviderOptionsResponse types
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
For replacing a single cue's voice (e.g., French Canadian → France French female)
without re-running the full pipeline.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Whisper's snap_pause_point() finds the nearest sentence boundary
independently per cue, which can move a later cue's pause_point before
an earlier cue's. The renderer then sorts by pause_point, producing
non-sequential cue indices in the timeline.
Add a forward monotonicity pass (clamp each pause_point >= previous) at
three layers for defense-in-depth:
- whisper_service: Phase 3 after consolidation
- video_renderer: before temporal sort in _render_pause_insert_method
- rerender_accessible_video: in _build_placements_with_adjustments
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Allow production/admin users to move jobs back to pending_qc from
completed, pending_final_review, rejected, qc_feedback, tts_failed,
render_failed, approved_english, and approved_source statuses. Includes
single-job endpoint, bulk endpoint, JobDetail inline form with required
notes, bulk action in JobsList with confirmation modal, and a Review
Notes card on the job overview tab.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Changed useEffect dependency from full pausePoint object to just
cue_index. This prevents the input from resetting when parent re-renders
cause the pausePoint object reference to change while editing the same
pause point.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
The input was reformatting with .toFixed(3) on every keystroke, causing
backspace to appear to insert random digits. Changed to string-based
input state with conversion/validation only on blur or save.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
When an AD cue is deleted, all subsequent cues shift positions but their
MP3 files remain at the old indices. This adds handling to automatically
queue TTS regeneration for all cues that shifted after a deletion.
Changes:
- VttEditor: Add onCueDeleted callback to notify parent of deletions
- QCDetail: Track deletion context and queue TTS for all shifted cues
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
When a new AD cue is inserted in the middle of existing cues, the system
now automatically queues TTS regeneration for the new cue AND all cues
that shifted positions. This ensures MP3 file indices stay synchronized
with VTT cue indices, preventing cues from being silently dropped during
re-render.
Changes:
- VttEditor: Add onCueInserted callback to notify parent of insertions
- QCDetail: Track insertion context and queue TTS for all shifted cues
- rerender_accessible_video: Add warning log when cue/MP3 count mismatch
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
QC approval now transitions jobs directly to pending_final_review
since translation, TTS, and accessible video rendering happen before
QC review. Removes unnecessary translate_and_synthesize_task trigger
on approval.
- Update approve_source() to use PENDING_FINAL_REVIEW status
- Update bulk_approve_jobs() to use PENDING_FINAL_REVIEW status
- Remove translate_and_synthesize_task.delay() calls from both endpoints
- Update JobDetail progress indicator to reflect new flow
- Update CLAUDE.md state machine documentation
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Display 0-based cue index badges in the editable AD cue list to match
the numbering shown on the timeline preview, helping users associate
timeline markers with their corresponding editable cues.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Since accessible video is now rendered immediately on upload, the method
selection (pause_insert vs overlay) is moved from QC Review to the New
Job panel. The bulk approval modal for selecting the method is removed.
- Add method selector UI to NewJob.tsx below accessible video checkbox
- Remove method selector from QCDetail.tsx approval flow
- Remove bulk approval modal from QCList.tsx
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Add ad_cue_index as secondary sort key when sorting placements, ensuring
that consolidated cues maintain their original VTT order (cue 0 before cue 1).
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Add ability to save voice settings changes in QC Review screen without
needing to approve the job. When saved, all TTS segments are regenerated
across all languages with the new voice settings.
Changes:
- Add PUT /jobs/{id}/tts-preferences endpoint to update TTS preferences
- Add UpdateTTSPreferencesRequest schema
- Add updateTTSPreferences API method and useUpdateTTSPreferences hook
- Add Save Voice Settings button with change detection to QCDetail
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Eliminate the two-step save process for VTT cue edits. Previously users
had to (1) save individual cue edits, then (2) click "Save Changes" in
a separate yellow notification box. Now saving a cue immediately persists
to the database and queues TTS regeneration for AD cues.
Changes:
- Add onCueSave callback prop to VttEditor for immediate persistence
- Add per-cue saving indicators and error handling with retry
- Remove hasUnsavedChanges state and yellow "Unsaved Changes" box
- Remove Ctrl+S keyboard shortcut (no longer needed)
- AD cue saves automatically queue TTS regeneration
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Existing jobs in the database don't have source_ms field. Making it
optional allows the API to load these jobs without validation errors.
The re-render task already handles the fallback to original_ms when
source_ms is None.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
The re-render task was using pause point coordinates from the accessible
video timeline (which includes freeze frame durations) instead of the
original source video coordinates. This caused pause points to exceed
the source video duration and get clamped incorrectly.
Changes:
- Add source_ms field to PausePointData model to store source video cut point
- Update video_renderer.py to populate source_ms when building pause points
- Update rerender_accessible_video.py to use source_ms for placement calculations
- Apply user adjustments as relative offsets (delta-based adjustment)
- Update API responses and TypeScript types to include source_ms
- Add backward compatibility fallback for jobs without source_ms
Note: Existing jobs need to be re-processed from initial render to populate
the new source_ms field.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
The task was created but not imported in the Celery task registry,
causing "Received unregistered task" error when triggering re-render.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
The rendering_qc status was added to the Python model but was missing
from the MongoDB schema validator, causing WriteError when setting
job status during QC re-rendering.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
When the user edits the audio description VTT and saves, the system now:
1. Compares original vs current AD cue text
2. Identifies which cues were modified
3. Automatically queues TTS regeneration for modified cues
4. Updates the Render Changes panel to show queued regenerations
This enables the "Render Changes" button when AD script edits are saved,
ensuring the accessible video can be re-rendered with updated TTS audio.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Pause points were being stored with source video timestamps instead of
rendered video timeline coordinates. This caused misalignment between
the pause point markers and freeze frame segments in the timeline UI.
Now pause points are calculated from the freeze frame segment start
positions in the rendered timeline, ensuring they align correctly
with the AD audio segments.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
The generate_signed_url() was called with expiration=3600 as an integer,
but GCS expects a datetime or timedelta. Now uses gcs_service.get_signed_url()
which properly calculates the expiration timestamp.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
The AccessibleVideoEditStateResponse schema expects string timestamps
but the API was passing raw datetime objects from MongoDB. Now converts
last_render_at and requested_at to ISO format strings.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Reorder workflow: translations now happen BEFORE QC Review step
- Add language tabs to switch between translated languages in QC
- Add video mode tabs (Original Video / Accessible Video)
- Add interactive timeline preview showing video segments and AD cues
- Enable pause point adjustment with millisecond precision
- Add TTS regeneration queue for selective cue re-synthesis
- Add re-render controls with optional Whisper refinement
- Persist video segments and TTS MP3s to GCS for editability
- Add new RENDERING_QC job status for re-render operations
- Create 5 new API endpoints for accessible video editing
- Add rerender_accessible_video.py Celery task
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
The previous implementation incorrectly used _get_video_duration which
in Cloud Run mode uses the cached source video URI instead of actually
measuring the freeze segment files. This caused all freeze segments to
report the source video duration (~78s) instead of their actual duration.
Changed to use _get_video_duration_local directly since freeze segments
are local files and need to be measured directly via ffprobe.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Subtitles were appearing progressively out of sync (~1.0s early per AD)
because the VTT retimer calculated freeze durations theoretically
rather than using actual rendered segment durations.
Changes:
- video_renderer: Measure actual freeze segment duration after creation
- video_renderer: Return updated placements with actual_freeze_duration
- vtt_retimer: Prefer actual_freeze_duration over calculated values
- render_task: Pass actual durations to VTT retimer
This ensures subtitle timing matches the real video timeline regardless
of any FFmpeg encoding variations.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
When TTS synthesis fails after 3 retries, the system now:
- Sends problematic cue text to Gemini for TTS-safe rewriting
- Updates the VTT file in GCS with rewritten text
- Retries TTS synthesis with the new text
- Records successful rewrites in job.tts_rewrites field
UI changes:
- JobDetail shows amber caution box with original/rewritten text
- JobsList shows warning icon next to jobs with rewrites
- Error display clarifies text shown is "after rewrite attempt"
Files changed:
- backend/app/models/job.py: Add tts_rewrites field
- backend/app/prompts/gemini_tts_rewrite.md: New prompt template
- backend/app/services/gemini.py: Add rewrite_tts_cue method
- backend/app/tasks/tts_synthesis.py: Add VTT update utilities
- backend/app/tasks/translate_and_synthesize.py: Rewrite+retry logic
- frontend/src/types/api.ts: Add TTSRewriteItem type
- frontend/src/routes/jobs/JobDetail.tsx: Caution display
- frontend/src/routes/jobs/JobsList.tsx: Warning indicator
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
MongoDB was rejecting status updates to 'tts_failed' and 'render_failed'
because these values weren't in the schema validator's enum, even though
they were defined in the Python JobStatus model.
This caused TTS failures to leave jobs stuck in 'tts_generating' status
with no error feedback to users - the WriteError from MongoDB prevented
the status and error fields from being updated.
The migration adds both failed statuses to the jobs collection validator.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Replace sequential browser-based bulk download with server-side zip
generation. When users select "Download All Files" from bulk actions,
the system now creates a single organized .zip file containing all
job assets.
Changes:
- Add POST /jobs/bulk/download endpoint that streams zip to client
- Add BulkDownloadRequest schema for the new endpoint
- Create zip_download.py service with streaming zip generation
- Update frontend to call new endpoint and download single zip file
- Organize files in zip by job title and language subdirectories
Zip structure:
accessible_video_YYYYMMDD_HHMMSS.zip
└── {job_title}/
├── source.mp4
└── {lang}/
├── captions.vtt
├── ad.vtt
├── ad.mp3
├── accessible_video.mp4
└── accessible_captions.vtt
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
The TTS error handling had a bug where failed jobs stayed in
'tts_generating' status instead of being set to 'tts_failed'.
Root cause: synthesize_cue_task used autoretry_for=(Exception,)
which raises the original exception after max retries, not
MaxRetriesExceededError. The exception handler never fired.
Changes:
- tts_synthesis.py: Replace autoretry_for with manual retry logic
that returns a failure dict on final failure instead of raising
- translate_and_synthesize.py: Add propagate=False to group.get()
to safely retrieve all results including failures
- translate_and_synthesize.py: Update outer exception handler to
set job status to tts_failed, store error details, and broadcast
status update via WebSocket
Now TTS failures will:
1. Set job status to 'tts_failed'
2. Store detailed error info (cue index, text, message)
3. Show error in UI with retry button
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Use typeof checks for proper type narrowing of unknown error fields
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Add POST /jobs/{id}/actions/retry_tts endpoint for retrying TTS
- Frontend shows TTS-specific error details (cue index, blocked text)
- Add "Retry TTS Generation" button on failed jobs
- Guides users to edit problematic AD text before retrying
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
