v1.4.0New
- MCP Server: connect HedyOS to Claude Code, Cursor, and other MCP-compatible editors
- Voice commands during calls — ask AI without switching windows
- Export transcripts as Markdown with speaker labels
Changelog
New features, improvements, and fixes in every release
v1.3.2Improvement
- Improved punctuation accuracy for Russian and English recognition
- Reduced memory usage during long recording sessions by 35%
- Faster model switching — no restart required
v1.3.0New
- Local Whisper support — fully offline transcription for sensitive meetings
- Automatic language detection across 5 new providers
- New keyboard shortcuts for start/stop and bookmark
v1.2.1Fix
- Fixed audio capture dropping out after macOS sleep/wake cycle
- Resolved rare crash when switching audio devices mid-session
- Corrected timestamp drift in sessions longer than 2 hours
v1.2.0New
- Real-time translation overlay — see translations as people speak
- Deepgram and AssemblyAI provider support
- Session search across all past transcripts
- Dark/light theme follows system appearance
v1.1.0Improvement
- Native ScreenCaptureKit integration — no virtual audio devices needed
- Improved speaker diarization accuracy
- Reduced app launch time by 40%