feat(transcribe): switch from OpenAI Whisper to WhisperX model

Replace openai/whisper with victor-upmeet/whisperx for 70x realtime
transcription with word-level timestamps and speaker diarization support.
Update audio input param from `audio` to `audio_file`, add diarization
config with HuggingFace token, and update output parsing for WhisperX
segment format with fallback for legacy Whisper output.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Author: aramb-dev

STUDIO_ROADMAP.md

@@ -0,0 +1,371 @@
+# Transcription Studio Roadmap
+
+> **Vision**: Transform Transcription Studio into a world-class, professional-grade audio transcription workspace that rivals dedicated desktop applications.
+
+---
+
+## Table of Contents
+
+1. [Phase 1: Foundation & Quick Wins](#phase-1-foundation--quick-wins)
+2. [Phase 2: Professional Audio Experience](#phase-2-professional-audio-experience)
+3. [Phase 3: Advanced Editing & Collaboration](#phase-3-advanced-editing--collaboration)
+4. [Phase 4: AI-Powered Features](#phase-4-ai-powered-features)
+5. [Phase 5: Enterprise & Scale](#phase-5-enterprise--scale)
+6. [Technical Debt & Infrastructure](#technical-debt--infrastructure)
+
+---
+
+## Phase 1: Foundation & Quick Wins
+
+**Timeline**: 1-2 weeks  
+**Goal**: Fix existing issues and establish a solid foundation
+
+### 1.1 Standalone Studio Page
+
+- [ ] Create `/studio` route as dedicated page (not just modal)
+- [ ] Deep linking support with session ID (`/studio?session=abc123`)
+- [ ] Browser history integration (back/forward navigation)
+- [ ] SEO meta tags for the studio page
+- [ ] Open Graph preview for shared links
+
+### 1.2 Fix Existing Bugs
+
+- [ ] **"Download All Formats" button** - Currently shows toast but does nothing
+- [ ] **DOCX export** - Currently exports plain text, not real DOCX format
+- [ ] **Audio URL persistence** - Improve localStorage handling for audio URLs
+- [ ] **Dark mode inconsistencies** - Fix contrast issues in segments view
+- [ ] **Native audio controls showing** - Hide native `<audio controls>` element
+
+### 1.3 Mobile Responsiveness
+
+- [ ] Responsive layout for tablets and phones
+- [ ] Stacked layout on mobile (audio player → transcript → controls)
+- [ ] Touch-friendly segment tapping
+- [ ] Swipe gestures for navigation
+- [ ] Bottom sheet for export options on mobile
+
+### 1.4 Keyboard Shortcuts
+
+| Shortcut | Action |
+|----------|--------|
+| `Space` | Play/Pause |
+| `←` / `→` | Skip -5s / +5s |
+| `Shift + ←` / `→` | Skip -30s / +30s |
+| `↑` / `↓` | Volume up/down |
+| `M` | Mute/Unmute |
+| `Ctrl/Cmd + C` | Copy transcript |
+| `Ctrl/Cmd + F` | Focus search |
+| `Escape` | Close modal / Clear search |
+| `1-9` | Jump to 10%-90% of audio |
+
+### 1.5 Loading & Empty States
+
+- [ ] Skeleton loaders for audio player
+- [ ] Skeleton loaders for transcript segments
+- [ ] Empty state illustrations
+- [ ] Error state with retry options
+
+---
+
+## Phase 2: Professional Audio Experience
+
+**Timeline**: 2-3 weeks  
+**Goal**: Create a best-in-class audio playback experience
+
+### 2.1 Advanced Audio Player
+
+- [ ] **Playback speed control** (0.5x, 0.75x, 1x, 1.25x, 1.5x, 2x)
+- [ ] **Loop selection** - Loop a specific time range
+- [ ] **A-B repeat** - Set start/end points for repetition
+- [ ] **Pitch correction** - Maintain pitch at different speeds
+- [ ] **Audio normalization** - Consistent volume levels
+
+### 2.2 Waveform Visualization
+
+- [ ] Real-time waveform display using Web Audio API
+- [ ] Zoomable waveform (pinch to zoom on mobile)
+- [ ] Click-to-seek on waveform
+- [ ] Segment regions highlighted on waveform
+- [ ] Current position indicator
+- [ ] Mini-map for long audio files
+
+### 2.3 Segment Navigation
+
+- [ ] Previous/Next segment buttons
+- [ ] Segment list with jump-to functionality
+- [ ] Auto-scroll transcript to current segment
+- [ ] Segment bookmarking
+- [ ] Quick navigation panel (timestamps sidebar)
+
+### 2.4 Audio Quality Enhancements
+
+- [ ] Noise reduction toggle (client-side)
+- [ ] Bass/Treble equalizer
+- [ ] Audio ducking for background music
+- [ ] Stereo/Mono toggle
+
+---
+
+## Phase 3: Advanced Editing & Collaboration
+
+**Timeline**: 3-4 weeks  
+**Goal**: Enable professional transcript editing workflows
+
+### 3.1 Inline Transcript Editing
+
+- [ ] Click-to-edit segment text
+- [ ] Real-time character count
+- [ ] Undo/Redo stack (Ctrl+Z / Ctrl+Y)
+- [ ] Edit history with timestamps
+- [ ] Diff view showing original vs edited
+- [ ] Batch find & replace
+
+### 3.2 Segment Management
+
+- [ ] Split segments at cursor position
+- [ ] Merge adjacent segments
+- [ ] Adjust segment timestamps manually
+- [ ] Delete segments
+- [ ] Add new segments
+- [ ] Drag-and-drop segment reordering
+
+### 3.3 Speaker Diarization UI
+
+- [ ] Visual speaker labels (Speaker 1, Speaker 2, etc.)
+- [ ] Custom speaker names (editable)
+- [ ] Color-coded speakers throughout transcript
+- [ ] Speaker timeline view
+- [ ] Filter transcript by speaker
+- [ ] Speaker statistics (word count, speaking time)
+
+### 3.4 Annotations & Comments
+
+- [ ] Add notes to specific timestamps
+- [ ] Highlight important sections
+- [ ] Tag segments (e.g., "action item", "question", "decision")
+- [ ] Export annotations separately
+- [ ] Comment threads on segments
+
+### 3.5 Version Control
+
+- [ ] Auto-save drafts to IndexedDB
+- [ ] Version history with restore
+- [ ] Compare versions side-by-side
+- [ ] Export specific versions
+
+---
+
+## Phase 4: AI-Powered Features
+
+**Timeline**: 4-6 weeks  
+**Goal**: Leverage AI to add intelligent features
+
+### 4.1 Smart Summarization
+
+- [ ] One-click transcript summary
+- [ ] Key points extraction
+- [ ] Action items detection
+- [ ] Meeting minutes generation
+- [ ] Custom summary length (brief/detailed)
+
+### 4.2 Translation
+
+- [ ] Translate transcript to 50+ languages
+- [ ] Side-by-side original + translation view
+- [ ] Export translated versions
+- [ ] Auto-detect source language
+
+### 4.3 Intelligent Search
+
+- [ ] Semantic search (find by meaning, not just keywords)
+- [ ] "Find similar segments"
+- [ ] Question answering ("What did they say about X?")
+- [ ] Topic clustering
+
+### 4.4 Auto-Correction
+
+- [ ] Grammar and spelling suggestions
+- [ ] Punctuation improvement
+- [ ] Filler word removal (um, uh, like)
+- [ ] Sentence boundary detection
+- [ ] Proper noun capitalization
+
+### 4.5 Content Analysis
+
+- [ ] Sentiment analysis per segment
+- [ ] Topic detection and tagging
+- [ ] Named entity recognition (people, places, organizations)
+- [ ] Keyword extraction
+- [ ] Word cloud generation
+
+### 4.6 Voice Commands
+
+- [ ] "Play", "Pause", "Skip forward"
+- [ ] "Go to minute 5"
+- [ ] "Find [keyword]"
+- [ ] "Summarize this"
+
+---
+
+## Phase 5: Enterprise & Scale
+
+**Timeline**: 6-8 weeks  
+**Goal**: Features for teams and power users
+
+### 5.1 User Accounts & Cloud Sync
+
+- [ ] User authentication (OAuth, email/password)
+- [ ] Cloud storage for transcriptions
+- [ ] Sync across devices
+- [ ] Transcription history dashboard
+- [ ] Usage analytics
+
+### 5.2 Team Collaboration
+
+- [ ] Shared workspaces
+- [ ] Real-time collaborative editing
+- [ ] Role-based permissions (viewer, editor, admin)
+- [ ] Assignment and task tracking
+- [ ] Activity feed
+
+### 5.3 Batch Processing
+
+- [ ] Upload multiple files at once
+- [ ] Queue management
+- [ ] Bulk export
+- [ ] Folder organization
+- [ ] Batch operations (delete, move, tag)
+
+### 5.4 Integrations
+
+- [ ] **Google Drive** - Import/export
+- [ ] **Dropbox** - Import/export
+- [ ] **Notion** - Export as page
+- [ ] **Slack** - Share transcripts
+- [ ] **Zapier/Make** - Automation workflows
+- [ ] **Zoom/Teams/Meet** - Direct recording import
+- [ ] **YouTube** - Transcribe from URL
+- [ ] **Podcast RSS** - Batch transcribe episodes
+
+### 5.5 API & Webhooks
+
+- [ ] Public REST API for transcriptions
+- [ ] Webhook notifications (transcription complete, etc.)
+- [ ] API key management
+- [ ] Rate limiting dashboard
+- [ ] SDK for common languages
+
+### 5.6 Advanced Export Options
+
+- [ ] **PDF** - Professional formatted document with timestamps
+- [ ] **DOCX** - Proper Word document with styles
+- [ ] **SRT/VTT** - Subtitle formats (already implemented)
+- [ ] **JSON** - Full data export with all metadata
+- [ ] **XML** - Structured export
+- [ ] **EDL** - Edit Decision List for video editors
+- [ ] **Markdown** - With timestamps and speaker labels
+- [ ] **HTML** - Interactive web page
+- [ ] **CSV** - Spreadsheet format
+
+---
+
+## Technical Debt & Infrastructure
+
+### Performance Optimizations
+
+- [ ] Virtualized segment list for long transcripts (react-window)
+- [ ] Lazy load audio waveform
+- [ ] Web Workers for audio processing
+- [ ] Service Worker for offline support
+- [ ] Optimize bundle size (code splitting)
+
+### Code Quality
+
+- [ ] Extract AudioPlayer into reusable component
+- [ ] Create custom hooks for audio state management
+- [ ] Add comprehensive unit tests
+- [ ] Add E2E tests with Playwright
+- [ ] Storybook for component documentation
+
+### Accessibility (a11y)
+
+- [ ] Full keyboard navigation
+- [ ] Screen reader support (ARIA labels)
+- [ ] High contrast mode
+- [ ] Reduced motion support
+- [ ] Focus indicators
+
+### Internationalization (i18n)
+
+- [ ] UI translation support
+- [ ] RTL language support
+- [ ] Locale-aware formatting (dates, numbers)
+
+---
+
+## Success Metrics
+
+| Metric | Target |
+|--------|--------|
+| Time to first transcription | < 30 seconds |
+| Studio load time | < 2 seconds |
+| Mobile usability score | > 90 |
+| Lighthouse performance | > 90 |
+| User satisfaction (NPS) | > 50 |
+| Export success rate | > 99% |
+| Audio playback reliability | > 99.5% |
+
+---
+
+## Priority Matrix
+
+```
+                    HIGH IMPACT
+                        │
+    ┌───────────────────┼───────────────────┐
+    │                   │                   │
+    │  • Standalone     │  • Waveform       │
+    │    page           │  • AI Summary     │
+    │  • Mobile         │  • Collaboration  │
+    │  • Keyboard       │  • Cloud sync     │
+    │    shortcuts      │                   │
+    │  • Fix exports    │                   │
+LOW ├───────────────────┼───────────────────┤ HIGH
+EFFORT                  │                   EFFORT
+    │                   │                   │
+    │  • Dark mode      │  • Voice commands │
+    │    fixes          │  • Video editor   │
+    │  • Loading        │    integration    │
+    │    states         │  • Real-time      │
+    │                   │    collab         │
+    │                   │                   │
+    └───────────────────┼───────────────────┘
+                        │
+                    LOW IMPACT
+```
+
+---
+
+## Getting Started
+
+**Recommended order of implementation:**
+
+1. **Week 1-2**: Phase 1 (Foundation) - Fix bugs, add standalone page, keyboard shortcuts
+2. **Week 3-4**: Phase 2.1-2.2 (Audio) - Playback speed, waveform visualization
+3. **Week 5-6**: Phase 3.1-3.3 (Editing) - Inline editing, speaker diarization
+4. **Week 7-8**: Phase 4.1-4.2 (AI) - Summarization, translation
+5. **Week 9+**: Phase 5 (Enterprise) - Based on user feedback and demand
+
+---
+
+## Notes
+
+- All features should maintain backward compatibility
+- Progressive enhancement - basic functionality works without JS
+- Privacy-first - no data sent to servers without explicit consent
+- Offline-capable where possible
+- Mobile-first responsive design
+
+---
+
+*Last updated: January 2026*