Repolode
Submit
cjpais

cjpais/Handy

A free, open source, offline speech-to-text desktop app that pastes transcribed words into any text field on Windows, macOS and Linux.

Open-source alternative to:Dragon NaturallySpeakingApple Dictation

View on GitHubWebsite
Stars
34,228
Forks
3,243
Open issues
183
Last commit
22 hours ago

What is it?

Handy is a cross-platform desktop app for speech transcription. You press a keyboard shortcut, speak, and the resulting text is pasted into whatever application has focus. Audio is processed on your own machine, so nothing is sent to the cloud.

Speech is filtered with Silero-based voice activity detection to skip silence. Transcription uses either Whisper models (Small, Medium, Turbo, Large), with GPU acceleration when available, or Parakeet V3, a CPU-oriented model with automatic language detection. The shortcut can be set to hold-to-record, toggle, or both.

Handy also offers a command-line interface for controlling a running instance, such as toggling recording from a window manager or hotkey daemon, plus a Raycast extension. On Linux, it relies on tools like xdotool, wtype or dotool for text input.

Key features

  • 01Runs speech recognition entirely offline on your computer.
  • 02Start and stop recording with a configurable global shortcut.
  • 03Supports Whisper models and the Parakeet V3 model, with custom Whisper GGML models.
  • 04Filters silence with Silero voice activity detection.
  • 05Available on Windows, macOS and Linux, with a command-line interface for remote control.

Use cases

  • Dictating messages, emails or documents into any app without sending audio to a cloud service.
  • Giving keyboard-free or reduced-typing input as an accessibility aid.
  • Hands-free note taking or drafting while working on a computer.
  • Integrating voice control into a Wayland or Sway/Hyprland workflow via CLI flags.

Quick start

  1. 1

    Download the latest release from the releases page

  2. 2

    On macOS, install via Homebrew cask

    brew install --cask handy
  3. 3

    Launch Handy and grant microphone and accessibility permissions

  4. 4

    Configure keyboard shortcuts in Settings and start transcribing

Things to know

  • Wayland support on Linux is limited and needs wtype or dotool for reliable text input.
  • Shortcuts using the fn (Globe) key only work on Apple keyboards.
  • The Homebrew cask and winget package are not maintained by the Handy developers.

Topics

#speech-to-text#offline#whisper#privacy#accessibility#cross-platform#tauri#rust#tauri-v2

FAQ

What is Handy?
A free, open source, offline speech-to-text desktop app that pastes transcribed words into any text field on Windows, macOS and Linux. Handy is a cross-platform desktop app for speech transcription. You press a keyboard shortcut, speak, and the resulting text is pasted into whatever application has focus. Audio is processed on your own machine, so nothing is sent to the cloud.
Is Handy free to use? What license does it have?
Handy is open source under the MIT license. Check the license terms before using it in commercial products.
Is Handy still maintained?
Actively maintained. The last commit to Handy was on Oct 20, 2026.
What can I use Handy for?
Dictating messages, emails or documents into any app without sending audio to a cloud service. Giving keyboard-free or reduced-typing input as an accessibility aid. Hands-free note taking or drafting while working on a computer. Integrating voice control into a Wayland or Sway/Hyprland workflow via CLI flags.
How do I install Handy?
Download the latest release from the releases page On macOS, install via Homebrew cask: brew install --cask handy Launch Handy and grant microphone and accessibility permissions Configure keyboard shortcuts in Settings and start transcribing
Where can I get Handy?
The source code and installation instructions are on GitHub: https://github.com/cjpais/Handy

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects