Local speech-to-text for macOS, Windows, and iOS.
On-device AI transcription - no cloud, no API keys, no subscriptions.
Your voice data never leaves your device.
| Platform | Status | Download | Repository |
|---|---|---|---|
| macOS | Available | Releases | typewhisper-mac |
| Windows | Available | Releases | typewhisper-win |
| iOS | Beta | TestFlight | typewhisper-ios |
TypeWhisper is a free, open-source speech-to-text app that runs entirely on your device. Press a hotkey, speak, and the transcription is pasted into whatever app you're using - no cloud round-trip required.
- 100% on-device - All AI processing happens locally. No telemetry, no data collection, no network requests during transcription.
- Multiple speech engines - Choose between WhisperKit, Parakeet TDT, Voxtral, Qwen3 ASR, Apple Speech (macOS), and ONNX-based engines (Windows). Cloud engines available via plugins.
- System-wide dictation - Global hotkey with push-to-talk, toggle, or hybrid mode. Works in any app.
- Per-app profiles - Automatically switch language, engine, and post-processing based on the active app or website.
- File transcription - Drag and drop audio/video files, export subtitles as SRT or WebVTT.
- Local HTTP API - Integrate with scripts, shortcuts, and automation tools.
- Plugin system - Extend with cloud providers (Groq, OpenAI, Gemini, Deepgram, AssemblyAI), actions, and custom add-ons.
- Dictionary and snippets - Custom terms, corrections, and text expansions with dynamic placeholders.
- iOS keyboard - Custom keyboard extension for voice input in any iOS app.
TypeWhisper is free and open source under GPLv3. If you find it useful, consider supporting development: