MongoDB 7.0 rejects the invalid key with code 28811, causing 500 on
GET /organizations/{id}/members.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
u.get("key", "") returns None when key exists with null value in MongoDB,
causing Pydantic ValidationError on MemberDetail.email/full_name: str → 500.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Cloud Run-generated freeze segments caused FFprobe to return code 1 with
empty stderr when dispatched to the Celery ffmpeg queue, crashing the
render for every language. The freeze segments are created to an exact
pre-computed duration (ad_duration + silence_before + silence_after),
so probing is unnecessary — assign that value directly instead.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Backend (Phase A):
- A1: Adaptive silence buffer — natural_gap_ms persisted per cue; renderer computes
per-cue silence_before/silence_after instead of fixed 500ms; per-cue silence files
- A2: Forward-preferred snap — snap_pause_point prefers boundaries up to 4s ahead
over boundaries within 1.5s behind, reducing mid-scene cuts
- A3: Min-gap validation — pause points with < 200ms gap trigger forward search
to the next acceptable gap
- natural_gap_ms added to PausePointData model and api.ts type
- New config fields: whisper_snap_forward_window, whisper_snap_backward_window,
ad_silence_buffer_default, ad_silence_buffer_min_after, ad_min_acceptable_gap
- Tests: test_whisper_snap.py (13 tests), test_video_renderer_buffers.py
Frontend (Phase B):
- B1: Drag pause-point markers — pointer state machine with 3px move threshold,
clamp to min/max bounds, click-without-move still opens PausePointEditor
- B2: Drag freeze blocks — orange blocks translate with linked pause point
- B3: Time tooltip visible during drag, hidden on release
- Tests: TimelinePreview.drag.test.tsx (10 tests)
Fixes:
- Share link pointed to ai-sandbox.oliver.solutions — added app_url to Settings
with correct optical-dev.oliver.solutions default; share_url now configurable
via APP_URL env var
- Removed all ai-sandbox.oliver.solutions references from docker-compose,
apache config, docs, and scripts
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
list_org_members and _membership_from_doc used bracket access on role_in_org
which raises KeyError if the field is absent (old docs or direct DB inserts).
Also handles ValueError if the stored value doesn't match a valid OrgRole.
Falls back to OrgRole.MEMBER in both cases.
Fixes 500 on GET /organizations/{org_id}/members.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- glossary_service: add step-by-step debug/warning logs at each early-return point so
the exact failure reason is visible in worker logs (project not found, no active version, etc.)
- glossary_service: guard against source_term_lower=None in ahocorasick automaton build
- glossary_service: guard against target_locale=None in _get_translation
- glossary_service: add full traceback to the outer exception catch for easier debugging
- JobsList: fix statusFilter stale state — useEffect now always syncs with URL params,
clearing the filter when no ?status= param is present (previously the filter was never
cleared, so navigating from /jobs?status=X to /jobs kept the old filter)
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
lameenc.encode() returns bytearray, but google-cloud-storage's
_to_bytes() only accepts bytes/str — causing TypeError on every
upload_from_string() call. Cast to bytes() before returning.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Gemini TTS _pcm_to_mp3 used pydub.AudioSegment.export(format='mp3') which
requires a system ffmpeg binary. Worker containers don't have ffmpeg installed
(video ops run on Cloud Run). Switch to lameenc which is pure Python and
encodes PCM→MP3 without any system binary.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- POST /{job_id}/actions/blocked_on_source (L-18): linguist/reviewer flags a source
video issue; moves job to QC_FEEDBACK and records blocked_on_source_reason/at/by
- POST /{job_id}/actions/promote_to_qc (PR-10): production/admin manually bypasses
AI processing for edge-case failures; adds audit history entry
- Reset reviewed_cues to 0 on submit_for_review (R-12) so reviewer must re-acknowledge
all cues after each linguist resubmit
- Add assert_job_in_user_org + get_user_org_ids to core/dependencies.py (used by
the new endpoints and the cross-tenant isolation test suite)
- Remove unused ingest_and_ai_task / translate_and_synthesize_task imports
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Blocks 1–5 of stabilization plan:
SECURITY
- validation.py: restore settings.upload_max_video_bytes (T-14 regression fix)
and JSON object key validation that was incorrectly removed
- MT-18: add accessible_org_ids filter to list_for_reviewer/list_for_linguist
so reviewers/linguists only see jobs from their own org in QC queue
- MT-17: add Membership.team_ids[], write to it on invitation acceptance and
direct team add/remove; migration backfills from Team.member_user_ids
- MT-19: validate all target_team_ids belong to invitation's org_id at creation
TESTS
- Restore test_cross_tenant_isolation.py (was deleted, only .pyc remained)
- Extend with MT-18 reviewer org isolation tests
QUICK WINS
- W-8: remove time.sleep(1) + dead debug block from POST /jobs (task was
undefined — would have caused NameError → HTTP 500 on every job creation)
- T-13: warn at startup when REDIS_URL configured but connection failed
- T-16: skip language_qc lifespan migration when count=0 (no DB scan on startup)
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Heavy pipeline tasks (ingest, translate, render, rerender) now dispatch
to a Cloud Run Job (va-worker) instead of local Celery workers. optical-dev
runs only api + lightweight worker (notify/embed) within its 2-CPU budget.
- backend/app/tasks/runner.py — Cloud Run Job entrypoint
- backend/app/services/cloud_run_dispatch.py — replaces .delay() for heavy tasks
- backend/Dockerfile.cloudrun — Cloud Run worker image (ffmpeg included)
- docker-compose.optical-dev.yml — 2-CPU safe overrides, disables heavy workers
- cloudbuild.yaml — builds va-worker image and updates Cloud Run Job
- deploy-dev.sh — uses 3-file compose, builds only api+worker locally
- routes_jobs, routes_admin_production, ingest_and_ai, translate_and_synthesize
— all dispatch sites updated to use cloud_run_dispatch.dispatch()
USE_CELERY_FALLBACK=true in .env.local to use Celery locally during dev.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Prevent PM in org A from assigning linguist/reviewer from org B.
Added _assert_user_in_job_org() helper that resolves job org_id (with
project fallback) and checks db.memberships for the assignee. Also added
assert_user_in_org() and get_job_or_403() to core/authz.py for use in
upcoming MT-13 and MT-15 commits.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
W-4: team assignment (linguist/reviewer) stored on job at creation,
auto-assigned to all language QC states on first GET /language-qc
(lazy init via auto_assign_defaults)
L-3 WS: broadcast_to_job when reviewer opens VTT for editing;
QCDetail shows "User X is editing [lang]" banner (auto-clears 5s)
R-5: comment broadcast via broadcast_to_job on add_comment();
QCDetail invalidates comments query on language_qc_comment WS event
L-15: QCDetail subscribes to language_qc_assigned WS event →
refetches lang-qc data and shows toast
R-7: VttEditor gets onCuePlay prop; AD editor in QCDetail wires
handleAdCuePlay → switches to accessible video mode, seeks & plays
T-15: beforeunload warning in NewJob while upload is in progress
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Changes:
- Dashboard: add project_manager case (final review / QC counts / new job widgets)
and production case (AI pipeline / failures widgets)
- Sidebar: add project_manager to Final Review and Audit Log nav items;
live badge counts for QC Queue (pending_qc) and Final Review (pending_final_review)
- App.tsx: add project_manager to Final Review and Audit Log RoleGates (W-10, PM-18)
- Login: role-based redirect after login — linguist/reviewer → /qc/queue, others → /
- language_qc._assert_can_approve: enforce two-stage QC; remove linguist self-approve
fallback; require reviewer assignment + submitted_for_review_at (W-6)
- routes_jobs.complete_job: allow project_manager to complete jobs (W-9)
- notify.py: re-enable email notifications (W-7)
- Fix 400 on cue save: treat empty-string audio_description_vtt/captions_vtt as absent
both in backend (truthy check) and frontend (|| undefined) — root cause was adVtt
initialising to '' when job has no AD track
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Two-stage QC workflow: linguist edits + submits → reviewer approves/rejects per language.
New statuses: in_progress, pending_review, in_review. New service functions: submit_for_review,
open_review, assign_reviewer, reassign_reviewer, add_comment. Linguist and reviewer deadlines.
- Reject now resets language to in_progress so linguist can iterate without full re-assignment.
- QC comment threads per language (append-only), visible to all assignees.
- Email notifications via Mailgun on: assignment, submit-for-review, comment, approve, reject.
Best-effort (failures do not roll back QC actions). asyncio.gather for parallel fan-out.
- New audit actions: LANGUAGE_QC_REVIEWER_ASSIGN/REASSIGN, LANGUAGE_QC_SUBMIT,
LANGUAGE_QC_OPEN_REVIEW, LANGUAGE_QC_COMMENT.
- Inline project picker in NewJob: "＋ Create new project…" option with name, default
languages, default linguist, default reviewer. Pre-fills languages on the new job.
- Project model extended with default_languages, default_linguist_id, default_reviewer_id.
- RBAC: CLIENT org-members can now create projects (backend guard relaxed).
- LinguistQueue: role toggle "As linguist / As reviewer" + new status tabs.
- QCDetail: two-slot assignment cards (linguist + reviewer), deadline display, role-aware
action buttons, comments panel with optimistic insert and 15s refetch.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
text-embedding-004 and text-multilingual-embedding-002 are not available
through this API key. gemini-embedding-001 (768-dim, multilingual) is.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Projected docs only have _id/source_term/translations; validating against
GlossaryTerm (which requires glossary_id, version_id, source_term_lower)
caused 500 on the terms endpoint. Return plain dicts instead.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
PyMongo Collection raises NotImplementedError on bool(), so 'if not self.collection'
crashes on every audit log write. Changed to 'if self.collection is None'.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- login and microsoft_login routes now use Depends(get_database) instead
of creating a per-request MongoClient — removes connection-pool churn
under load
- MicrosoftAuthService._get_openid_config/_get_jwks/validate_token are
now async, using httpx.AsyncClient instead of blocking requests.get —
removes ~400ms event-loop block per Microsoft login
- Removed unused AsyncIOMotorClient import from routes_auth.py
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
user.role stored as plain string in MongoDB — calling .value on it
caused AttributeError on every login, blocking all auth.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Job.language_qc dict tracks per-language status (pending/in_review/approved/rejected)
with full event history; qc_assignments denormalized array enables efficient queue queries
- language_qc service handles assign/reassign/approve/reject/reopen with atomic DB updates,
audit logging, and auto-advancement to pending_final_review when all languages approved
- Linguists can only edit VTT and trigger re-renders for their assigned language (403 guard)
- return_to_qc resets all language statuses while preserving assignments
- routes_language_qc.py: 7 new endpoints; /me/language-qc-queue for linguist queue
- Startup migration idempotently seeds language_qc for all existing jobs
- Frontend: LanguageQCState types, API methods, LinguistQueue page, QCDetail redesigned
with per-language status badges, assignment dropdown, inline approve/reject buttons,
progress bar, and reject modal; My QC Queue sidebar link
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
authz.py (new):
- MembershipContext — per-request membership dict for the current user
- get_membership_context FastAPI dependency
- require_org_role(min_role) — dependency factory keyed off org_id path param
- require_platform_admin()
- OrgScopedQuery — adds organization_id filter; platform admin passes through
- bump_user_membership_cache — invalidates Redis key on membership writes
dependencies.py:
- get_accessible_project_ids now queries memberships collection first;
legacy pm_client_ids / team.member_user_ids fallback retained until migration runs
(four job-route access checks at lines 608/1054/1181/1538 are fixed via this function)
routes_clients.py:
- _assert_pm_or_admin and _assert_client_access are now async and query memberships
- All 10 call sites updated with await + db arg
emailer.py:
- Switched from SendGrid to Mailgun REST API via httpx (already in requirements)
- _send() is now fully async; same public method signatures preserved
- send_completion_email uses _send()
config.py:
- Added mailgun_api_key, mailgun_domain, mailgun_from settings
- sendgrid_api_key kept with empty default for backward compat
migration_2026-04-28-000003:
- Backfills job.organization_id from project.client_id
- Creates (organization_id, status, created_at) sparse index on jobs
routes_organizations.py / routes_invitations.py:
- Call bump_user_membership_cache after every membership write
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Backend:
- models/invitation.py — Invitation model + create/accept/preview schemas
- routes_invitations.py — org-scoped POST/GET/DELETE + public preview/accept endpoints
Single-use token via find_one_and_update; sha256(token) stored in DB, plaintext in email URL
- emailer.py — _send() helper; send_invitation_email, send_welcome_email, send_password_reset_email
send_completion_email refactored to use _send()
- migration_2026-04-28-000002 — creates invitations collection with TTL index (30d audit trail)
- routes_auth.py — new MS SSO users provisioned with zero memberships instead of role=PRODUCTION;
they land on "no access" page until an admin invites them
- main.py — registers invitations_org_router and invitations_router
Frontend:
- routes/AcceptInvite.tsx — public page at /accept-invite?token=...
Four states: new user (name+password), existing user (confirm), MS user, already-member
- App.tsx — /accept-invite route outside RequireAuth
- types/api.ts — Invitation, InvitationCreate, InvitationPreview, InvitationAcceptRequest/Response
- lib/api.ts — listInvitations, createInvitation, revokeInvitation, previewInvitation, acceptInvitation
- hooks/useClients.ts — useInvitations, useCreateInvitation, useRevokeInvitation
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- New services/cost_tracker.py: sync httpx preflight()/record() + async wrappers;
BudgetExceeded exception; no-op when COST_TRACKER_BASE_URL is empty
- Preflight budget check added before ingestion (Gemini), per-language translation
(video-native + traditional), and per-language TTS dispatch
- _record_gemini_usage and _record_tts_cost now call cost_tracker directly;
removes broken asyncio.get_event_loop() hack from sync Celery worker
- Fix: _cost_ctx now threaded into extract_accessibility_targeted (video-native path)
- Fix: user_id/cost_project_id now propagated through dispatch_language_tts →
synthesize_cue_task.s() and the rerender_accessible_video.py re-render path
- Remove oliver-cost-tracker SDK dependency (was commented-out/never installed)
- Drop cost_tracker_outbox_path setting and get_cost_tracker() factory
- Update COST_TRACKER_BASE_URL default to optical-dev.oliver.solutions in
.env.prod.example, docker-compose.yml, and all Cloud Run service yamls
- Cloud Run yamls use Secret Manager ref (cost-tracker-api-key) for the API key
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
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>
- 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>
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>
- 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>
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>
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>
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>
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>
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>
- 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>
