
Use the shortcut key to convert speech to text: "Handy" by 36k star completely removes transcription from the cloud.
Handy: An open-source offline transcription tool. Speak using a shortcut key, and Whisper or Parakeet will transcribe the audio locally. The transcribed text can then be directly pasted into any application, with the audio remaining on your device.
Related tags
When it comes to voice-to-text, most people’s first reaction is to open a webpage for a cloud service – then upload the words they want to say, along with their privacy information, to someone else’s server. However, the open-source project “Handy,” with 36,618 stars, offers a different solution: press a shortcut key, speak, release your hand, and the text appears directly in your input field. The entire process is completed locally, without any audio leaving your computer. This desktop application, written in Rust, has gained another 339 stars today, consistently ranking among the top trending projects on GitHub.
Why Handy?
Handy’s positioning is very clear: a free, open-source, and extensible offline voice-to-text tool. The project provides four reasons on its website, handy.computer: free (essential tools shouldn’t be blocked by paywalls), open source (allowing users to extend and contribute), privacy (audio remains local), and simple (a tool that does one thing).
The workflow is compressed into four steps: press a configurable shortcut key (supporting continuous speech, single-tap switching, or using only one mode), speak, release your hand, and the transcribed text is directly pasted into any application you are currently using. There’s no need to switch windows or wait for uploads.
Technology Stack: Silero Noise Reduction, Whisper or Parakeet Transcription
Handy performs all processing locally. The recorded audio is first filtered by Silero’s VAD (Voice Activity Detection) to remove silent segments, and then passed to the transcription model. You can choose from:
- Whisper series (Small / Medium / Turbo / Large), which automatically accelerates on GPU;
- Parakeet V3, a CPU-optimized model with excellent performance and automatic language detection – even machines without dedicated graphics can use it smoothly.
It supports Windows, macOS, and Linux.
How to Use It
Installation is quite smooth: macOS users can use brew install --cask handy, while Windows users can use winget install cjpais.Handy (both packages are maintained by the community). For Debian/Ubuntu, it’s recommended to use apt install ./Handy_*.deb to automatically install all dependencies. After installation, simply launch the application, grant microphone and accessibility permissions, and configure the shortcut keys in the settings – you can start using it immediately.
In a time when cloud-based dictation services charge monthly fees and privacy incidents are frequent, an “offline” open-source solution with over 36,000 stars is not surprising. If you need to type a lot every day and don’t want to share your voice data in the cloud, Handy is definitely worth trying out.
Repository address: https://github.com/cjpais/Handy
More News
Want to see more quality content?
Browse all news