Handy
Handy is a free, open-source desktop app that transcribes your speech offline and pastes the text into whichever app you are using.
- GitHub stars
- 33k
- Last commit
- yesterday
- Latest release
- v0.9.7
- Licence
- MIT

Handy is a cross-platform desktop application for speech-to-text. You press a configurable keyboard shortcut, speak, and the transcribed words appear in the text field you are working in. Everything happens on your own computer, so audio is never sent to a cloud service. It is written in Rust with Tauri, released under the MIT license, and runs on Windows, macOS and Linux.
Recording can be set to hold-to-talk, toggle, or both. Silence is filtered with Silero voice activity detection, and transcription uses a model of your choice: Whisper models in several sizes with GPU acceleration when available, or the CPU-optimized Parakeet V3 with automatic language detection. The project states its aims as free, open source, private and simple, treating accessibility tooling as something that should not sit behind a paywall. Installers are on the releases page and website, with Homebrew available on macOS.
Key features
- Offline speech-to-text transcription
- Configurable hold or toggle keyboard shortcut
- Whisper models with GPU acceleration
- Parakeet V3 model with language detection
- Silero voice activity detection filters silence
- Pastes text into any app you use
Pricing: Free and open source under the MIT license.
