- Voice clone integration: When user selects voice clone in Write phase,
backend uses their uploaded voice sample + scene script text to generate
audio via qwen3/minimax/cosyvoice voice clone APIs
- Multi-tenant workspace storage: All podcast assets (audio, video, images,
charts) now use workspace-specific directories per user
- Chart preview improvements: Card-based B-Roll charts UI with thumbnails,
takeaway text, and action buttons; public endpoint for image serving
- Voice clone caching: In-memory LRU cache for voice samples (avoids
re-downloading per scene); frontend caches voice clone metadata
- Thread pool for voice clone: Audio generation uses ThreadPoolExecutor to
avoid blocking the FastAPI event loop
- Auto-detect voice clone IDs (vc_*, MY_VOICE_CLONE) to route correctly
- DB fallback for voice sample URL: Fetches from ContentAsset if not passed
- Fixed API URL resolution for chart previews
- Fixed GlassyCard DOM warnings for motion props
- Fixed ScriptGenerationProgressView syntax error
- Fixed usePodcastWorkflow scriptData reference
- Updated canSubmit to require all fields: topic, presenter avatar, voice selection, duration, speakers, and podcast mode
- Removed audio_only exception - avatar now required for all modes
- Updated tooltip message in CreateActions
- Add hamburger menu to Header with gradient styling
- Move Help, My Episodes, My Projects, New Episode into dropdown menu
- Move PodcastBiblePanel into AnalysisPanel header as icon button
- Display Bible details in a styled Popover
- Improve overall header UX and mobile responsiveness
- Add isGeneratingRef to track if generation is in progress
- Guard against double calls by checking ref before starting
- Reset ref in finally block to allow subsequent generations
- This fixes the issue where script generation was called twice in succession
- Wrap all AvatarSelector callback handlers in useCallback in CreateModal.tsx
- Add isCancelled flag pattern to RobustCamera useEffect
- Inline camera initialization to avoid stale closure issues
- Add proper cleanup on component unmount
- Ensure camera stream is properly stopped if component unmounts during initialization
- Remove unused initializeCamera function
- Add React.memo to CameraSelfie to prevent unnecessary re-renders
- Memoize callbacks in CameraSelfie
- Track previous open/facingMode state in RobustCamera to detect actual changes
- Add streamAttachedRef to prevent duplicate stream attachments
- Fix useEffect dependencies to prevent cleanup on parent re-renders
- Ensure camera only initializes on actual dialog open (not parent re-render)
- Consolidate stream attachment into single useEffect
- Remove race conditions from multiple competing effects
- Add proper cleanup with video element reset
- Simplify state management using single stream state
- Add isMountedRef to prevent state updates after unmount
- Improve error handling with specific error messages
- Add canvas flip correction for front camera mirror effect
- Use explicit MIME type for MediaRecorder (audio/webm;codecs=opus)
- Add error handling for audio playback
- Copy chunks before creating blob to prevent race conditions
- Add key prop to audio elements for proper re-rendering
- Add TextToSpeechButton to Outline, Takeaways, and Guest tabs in analysis phase
- Add help icon with tooltip to Research Queries explaining the workflow
- Change Run Research button to show 'Next: Select Query' when disabled
- Add hint text 'Select a query to continue' when no queries selected
- Make header stack responsive with column layout on mobile
- Add responsive breakpoints for tab sizes and padding
- Fix image preview max widths for mobile screens
- Add responsive font sizes for info boxes
- Adjust container padding for smaller screens
- Fix icon sizes for mobile devices
- Remove Research Queries section from AnalysisPanel.tsx
- Keep QuerySelection component as single source for research queries
- Remove unused props (onRunResearch, isResearchRunning, selectedQueries, etc.)
- Add pandas to requirements-podcast.txt for usage tracking
- Fix LLM prompt to return plain strings instead of objects for enhanced_ideas
- Add object-to-string normalization for LLM responses that return objects
- Skip bible generation in podcast-only mode (onboarding disabled)
- Skip alerts polling in AlertsBadge when in podcast-only demo mode
- AlertsBadge: Skip agent alerts fetch in podcast mode
- UserBadge: Skip system status fetch in podcast mode
- SystemStatusIndicator: Skip monitoring stats in podcast mode
This prevents 404 errors when frontend calls endpoints that don't exist in podcast-only demo mode.
- Avatar images now use full available width (max 280px, responsive)
- Auto-collapse info banner after 8 seconds
- Add 'Show tips' link to expand collapsed info
- Fix image sizing to use contain instead of cover for better visibility
- canSubmit now checks for avatar presence (uploaded, brand, or generated)
- Checks avatarFile, avatarUrl, avatarPreview, brandAvatarFromDb, brandAvatarBlobUrl
- Updated tooltip to reflect new requirement
Backend:
- Remove all legacy env var fallbacks (ALWRITY_FEATURE_PROFILE, ALWRITY_ROUTER_PROFILE, etc)
- Remove get_active_profile() from start_alwrity_backend.py
- Remove _env_flag_enabled() from app.py
- Use ALWRITY_ENABLED_FEATURES as single source of truth
Frontend:
- demoMode.ts now uses only REACT_APP_ENABLED_FEATURES
- Removed all legacy fallback keys (app_mode, demo_mode, podcast_only_demo_mode)
Usage:
ALWRITY_ENABLED_FEATURES=podcast # Podcast only
ALWRITY_ENABLED_FEATURES=all # All features (default)
- Remove fallback to ALWRITY_FEATURE_PROFILE, ALWRITY_ROUTER_PROFILE
- Primary env var is now ALWRITY_ENABLED_FEATURES (backend)
- Primary env var is REACT_APP_ENABLED_FEATURES (frontend)
- Add deprecation comments to all get_enabled_features() functions
- Update demoMode.ts with clear deprecation notes
Usage:
ALWRITY_ENABLED_FEATURES=podcast # Podcast only
ALWRITY_ENABLED_FEATURES=all # All features (default)
