- Translate to English setting: Added automatic translation of speech to English
- Settings refactored into React hooks for better state management
- Audio device switching capability
- Hysteresis to VAD (Voice Activity Detection) for more stable recording
- Major audio backend refactor for improved performance and reliability
- Moved audio toolkit into src-tauri directory for better permissions handling
- Model files no longer need to be downloaded separately for releases
- Updated settings components and transcription logic
- Audio toolkit permissions issues
- Various stability improvements
- Keycode bug that was causing input issues
- Whisper model optimization: switched to unquantized Whisper Turbo, updated Whisper Medium quantization to 4_1
- Removed 50ms delay feature flag for Windows (now applies to all platforms for consistency)
- Ctrl+Space key binding for Windows platform
- Windows crash issue
- Model loading on startup when available
- Windows paste functionality bug
- Microphone activation on demand: More efficient resource usage
- Less permissive VAD settings for better accuracy
- Improved microphone management and activation system
- Multiple models support: Users can now select from different transcription models
- Model selection onboarding flow
- Cleanup and refactoring of model management
- Enhanced user experience with model selection interface
- Better language and UI tweaks
- Different start and stop recording sounds: Enhanced audio feedback
- Recording sound samples for better user experience
- Build issues
- Auto-update functionality improvements
- Paste functionality using enigo library for better cross-platform compatibility
- Auto-update functionality: Application can now automatically update itself
- Footer displaying current version
- Improved menu system
- Better user interface for version management
- Enhanced update workflow
- Comprehensive build system: Support for Windows, macOS, and Linux
- Windows code signing for trusted installation
- Ubuntu/Linux build support with Vulkan
- Model file download and packaging for releases
- GitHub Actions CI/CD workflow
- Improved build process and release workflow
- Better cross-platform compatibility
- Various build-related issues across platforms
- Initial release of Handy
- Basic speech-to-text transcription functionality
- Voice Activity Detection (VAD) for automatic recording
- Cross-platform support (macOS, Windows, Linux)
- Tauri-based desktop application with React frontend
- Global keyboard shortcuts for activation
- Clipboard integration for automatic text insertion
- LLM integration for enhanced transcription processing
- Configurable settings including:
- Custom key bindings
- Audio device selection
- Microphone settings
- Push-to-talk functionality
- System tray integration with recording indicators
- Accessibility permissions handling for macOS
- Settings persistence with unified settings store
- Background operation capability
- Multiple audio format support with on-the-fly resampling
- Whisper model integration for high-quality transcription
- MIT License for open-source distribution
- Built with Tauri (Rust backend) and React (TypeScript frontend)
- Audio processing with cpal and whisper-rs
- Real-time transcription with performance optimizations
- Cross-platform keyboard event handling
- Modular architecture with managers for audio, models, and transcription