Handy

34.6k GitHub starscjpais/Handy

Free, private voice typing for your computer. Press a shortcut, speak, and your words appear in any app. Works offline on Mac, Windows and Linux.

Why we picked it

Handy turns your voice into typing anywhere on your computer. Press a shortcut, talk, let go, and the text lands in whatever box you're typing in: an email, a chat, a document or an AI prompt. The speech recognition runs on your own computer, so your voice is never uploaded and it keeps working offline. It's free, MIT licensed, works on Mac, Windows and Linux, and does one job well, which makes it a simple open source alternative to paid dictation apps.

About Handy

Handy is a free, open source app that turns your voice into text anywhere on your computer. Press a keyboard shortcut, say what you want to write, and Handy types it into whatever app you're using. The speech recognition runs on your own computer, so your voice is never sent to the cloud and it works without internet. It runs on Mac, Windows and Linux, and was made by independent developer CJ Pais.

What it can do

  • Type with your voice, anywhere. Your words are pasted straight into the text box you're in: email, chat, documents, a browser or an AI prompt.
  • Hold or tap to talk. Hold the shortcut while you speak and let go to finish, or tap once to start and again to stop. You choose the shortcut.
  • Private and offline. Recognition happens on your computer, so your recordings stay with you and it works without a connection.
  • Pick your speech model. Start with the recommended Parakeet model, which runs well on an ordinary processor, or choose OpenAI's Whisper models, from small and fast to large and accurate, which cover the most languages and use your graphics chip when there is one. Tiny Moonshine models suit older computers.
  • Spells things your way. Add custom words like names and jargon, remove filler words like "um", and look back through your recent transcripts.
  • Optional AI cleanup. An experimental setting tidies your words with AI before pasting, using OpenAI, Anthropic, Groq, OpenRouter, Apple Intelligence or a local model through Ollama.
  • Works with Raycast. A community extension lets you start recording, browse past transcripts and switch models from Raycast on Mac.

Handy, built-in dictation or Wispr Flow?

All three let you type by talking. They differ in privacy, polish and cost.

HandyThis page

Best forPrivate, free voice typing on any computer

Works in any app with one shortcut, and your voice never leaves your computer. Choose the speech model yourself. Plain dictation, without AI rewriting by default.

How you use it
Mac, Windows and Linux app
Open source
Yes (MIT)
Cost
Free
Built-in dictation

Best forQuick dictation with nothing to install

The dictation that comes with macOS and Windows. Already there and easy to start, but less accurate for long passages, and depending on your system and settings some audio may be processed online.

How you use it
Part of macOS and Windows
Open source
No
Cost
Free
Wispr Flow

Best forPolished dictation that tidies your words

A popular paid dictation app that cleans up filler words and formats your text with AI. Your audio is processed in the cloud.

How you use it
Desktop and phone apps
Open source
No
Cost
Free with limits; paid plans

A quick way to choose: just need a sentence now and then? Your computer's built-in dictation is fine. Want AI to polish what you say and don't mind paying? Wispr Flow. Want accurate voice typing that's free and stays private on your own computer? Use Handy.

What computer you need

Handy doesn't publish official minimum specs. As a guide, based on the size of each model:

  • Most laptops (8 GB of RAM, no graphics card): the recommended Parakeet model or Whisper Small, about 0.5 to 0.7 GB.
  • Older computers (4 GB of RAM): the tiny Moonshine models, some under 100 MB.
  • Best accuracy (16 GB of RAM and a graphics chip): Whisper Turbo or Large, 1 to 1.6 GB. Handy uses Metal on Apple Silicon Macs and Vulkan on Windows and Linux, which works with NVIDIA, AMD and Intel graphics.

How to get it

  • Easiest: download it from handy.computer or the releases page on GitHub, install it, and allow microphone and accessibility access when asked.
  • Mac: also available with Homebrew: brew install --cask handy
  • Windows: also available with winget: winget install cjpais.Handy
  • Linux (Debian, Ubuntu): install the downloaded file with sudo apt install ./Handy_*.deb so its dependencies come with it.

Good to know

  • The first time you pick a model, Handy downloads it. After that everything runs offline.
  • The recommended Parakeet model is English only. If you speak another language, pick a model that supports it: the Whisper models cover the most.
  • On Mac, shortcuts using the fn (Globe) key only work on Apple keyboards. Use ctrl, option, shift or command if you switch to another keyboard.
  • On Linux, typing works best with a helper tool installed (xdotool on X11, wtype or dotool on Wayland), and Wayland support is still limited.
  • The Homebrew and winget packages are kept up by the community, not the Handy developers.
  • Licence: MIT. Free to use, change and include in your own products, even paid ones, as long as you keep the licence notice.

What you can build with it

  • Dictating emails, messages and documents instead of typing
  • Talking to ChatGPT, Claude or coding assistants by voice
  • Voice typing for people with RSI or other accessibility needs
  • Private dictation that never sends audio to the cloud
  • Capturing quick notes and ideas hands-free

Comments

No comments yet. Be the first to share your thoughts.