Handless

Free, open-source macOS speech-to-text.
Press a shortcut, speak, get text in any app. Run locally for privacy or use cloud APIs.

简体中文 · 繁體中文 · 日本語

Features

• Local transcription -- download models in Settings, runs fully on-device
• Cloud STT via OpenAI or Soniox
• Voice Activity Detection (local models only)
• LLM post-processing to clean up or reformat transcriptions
• macOS (Intel & Apple Silicon)
• 17 languages

Install

Download for macOS (Intel & Apple Silicon)

Also available at handless.elwin.cc. Build from source: see BUILD.md.

CLI

Remote control (talks to a running instance):

handless --toggle-transcription    # Toggle recording
handless --toggle-post-process     # Toggle recording + post-processing
handless --cancel                  # Cancel current operation

Startup flags:

handless --start-hidden            # No main window
handless --no-tray                 # No tray icon
handless --debug                   # Verbose logging
handless --help                    # All flags

Combine freely: handless --start-hidden --no-tray

macOS: invoke the binary directly: /Applications/Handless.app/Contents/MacOS/Handless --toggle-transcription

Troubleshooting

Cmd+Shift+D opens the debug panel.

Contributing

See CONTRIBUTING.md. For translations: CONTRIBUTING_TRANSLATIONS.md.

License

MIT

Acknowledgments

Forked from Handy v0.7.8.

Whisper | whisper.cpp | NeMo Parakeet | Moonshine | SenseVoice | Silero VAD | Tauri | Handy