What is the Handy dictation app?

Handy is a free, open-source speech-to-text app that runs on your own computer. The project describes itself as ¹. Its README says the transcription happens on your computer, without sending any information to the cloud ². The code is public on GitHub under the MIT License, so you can read it, change it and reuse it. This page answers what Handy is and how to decide whether to try it. The open-source dictation guide compares it with the other free projects, and the dictation software overview covers the paid apps.

You press a shortcut, speak, and the words are pasted

The README gives the whole loop in three steps. You speak while the shortcut is active. You release it, and Handy processes the speech. The text is then pasted into the app you are using. In one sentence, you press a shortcut, speak, and have your words appear in any text field ³. The README says Handy listens while the shortcut is active, so you decide when it records.

The speech models are Whisper and Parakeet, in several sizes

Handy's README lists two families of speech model. Whisper comes in four sizes, Small, Medium, Turbo and Large, and it uses the graphics processor when one is available ⁴. The other family is Parakeet V3, which the README describes as a CPU-optimized model with automatic language detection ⁵. You download the models you want. Neither description gives a measured accuracy, so test each model on your own speech before you choose one.

Handy runs on Windows, macOS and Linux, and Linux needs a typing tool

Handy is listed as working on Windows, macOS and Linux ⁶. Linux takes more setup. The README asks you to install a typing tool that matches your display server: xdotool on X11, or wtype on Wayland. It also says that Ubuntu 26.04 uses Wayland by default, that wtype does not work there, and that you need ydotool with its systemd setup instead. Read that section before you install Handy on a current Ubuntu desktop.

The MIT License lets you read, change and reuse the code

Handy is released under the MIT License ⁷. That license lets anyone use, copy and modify the code, including for commercial use, as long as the copyright notice is kept. Open source here means the code is public and reusable. It does not by itself tell you whether the project has a support team or a release schedule.

Check the latest release tag before you download

On 16 October 2026 the GitHub releases page marked v0.9.8 as the Latest release ⁸. The release date was not shown as visible text on that page, so the tag is the only dated detail we confirmed. Download only from the project's own releases page at github.com/cjpais/Handy, and not from a copy on a third-party site.

Handy fits a single-purpose need, and the fit depends on your setup

Handy suits you if you want a free app that only transcribes, you can handle a setup step, and you want to read the code. It is a weaker fit if you need AI rewriting of your sentences, team management or a support contract. Handy also differs from the dictation built into your operating system, which needs no download. The free dictation software guide covers the built-in options, which need no download.

Two limits apply to every choice. First, we did not test Handy's accuracy on any microphone or room for this page, and a model's result depends heavily on both. Second, the claim that transcription stays on your computer comes from the project's own README. If network use matters to you, check it yourself. For the Mac-only alternative HEX, read what is the Hex dictation app. For a paid Mac app with a different license, read is VoiceInk free and open source.

Try Handy on your own speech in one sitting

1. Choose your platform, then download the release from the Handy releases page at github.com/cjpais/Handy.
2. Download the two model families you want to compare, Whisper and Parakeet V3.
3. Dictate the same paragraph with each model, and count the words you had to correct.
4. If privacy matters, watch the app's network use while you dictate, as the open-source guide describes.

For a general method that works with any app, see how to test dictation software. The dictation software entry in our glossary explains the term.