MightyMouse

Is there an open source Wispr Flow alternative?

Wispr Flow is good. It is also closed, and it sends your voice to a server. If you want to read the code behind the thing that hears you, you have options.

On your phone? Get the Mac app by email.

Short answer: Yes. MightyMouse is a free, MIT open source Wispr Flow alternative: hold a key, talk, and your words appear in any app, turned into text on your own computer, with the code public on GitHub at royherma/mightymouse.

Download free for Mac

The short version: MightyMouse types what you say into any app, free, with the voice turned into text on your own computer. Since 8 November 2026 the whole app is open source under the MIT license. You can read it, build it, change it and ship your own copy.

How voice typing works on your computer

You hold the fn key and talk. When you let go, your words land where your cursor is, in whatever app you are using. Slack, Mail, a browser box, a terminal.

The part that matters for privacy is where the speech becomes text. MightyMouse does that on your own computer with Whisper, an open speech model that runs locally. For the languages it supports it can also use NVIDIA's Parakeet model, which is faster. Your recording is not sent to a server to be turned into words. So it is free to run, because nobody has to pay for a server on our side, and it works with no internet once the model has downloaded.

Want the longer version? How to run Whisper locally on a Mac walks through the models, and dictating anywhere on your Mac covers everyday use.

How to install it, or build it yourself

Easiest: get the signed app from the MightyMouse home page. It updates itself. Sign-in uses a free MightyMouse account.

From source: you need a Mac and a current Node LTS release. Then run these four lines:

npm run dev starts the app. Use npm run build:mac when you want a packaged copy. An app you build yourself is not signed by Apple, so your computer asks you to confirm the first launch: right-click the app, then choose Open. The speech code lives in src/main/capture/whisper.ts if you want to make it faster or sharper. Pull requests are welcome.

MightyMouse vs Wispr Flow, side by side

Wispr Flow facts below come from its own pages as read in October 2026, so check them again before you decide. The full version is on the MightyMouse vs Wispr Flow page.

Other open source options

MightyMouse is not the only one. VoiceInk is a public Swift app for macOS under the GPL v3 license, and its repository describes itself as an open source alternative to Wispr Flow with no subscription. It is built for dictation alone. If dictation is the only thing you want, it is worth trying. See MightyMouse vs VoiceInk for the difference. A few smaller projects on GitHub target the same job. Check each one's license before you rely on it.

What MightyMouse adds on top of typing by voice: your phone shows your computer's live screen and becomes a trackpad, keyboard and microphone, and your coding agents can ask you questions that you answer from the lock screen.

What this costs you and where it falls short

Honest summary: pick Wispr Flow for polish across many devices and you are fine with the cloud. Pick MightyMouse or VoiceInk if you want code you can read and a voice that stays home.

Talk instead of typing, free, with code you can read.
No card to try it. Your voice stays on your computer.

Download free for Mac

Common questions

Is there an open source alternative to Wispr Flow?

Yes. MightyMouse is MIT licensed with its code on GitHub, and VoiceInk is a GPL v3 macOS app. Both turn your voice into text on your own computer instead of a server.

Is there a free open source dictation app for Mac?

MightyMouse voice typing is free with no word cap, and the app is open source under MIT. Hold fn, talk, and the words appear in any app. Sign-in uses a free MightyMouse account.

Is MightyMouse really open source?

Yes. The repository is public at github.com/royherma/mightymouse under the MIT license, which lets you use, change and redistribute it.

Does open source dictation work offline?

In MightyMouse, yes, once the speech model has downloaded. The speech is turned into text on your computer, so there is nothing to upload.

Is Wispr Flow open source?

We found no public source repository for it, and its privacy page says audio is sent to its servers for transcription. Check its site for the current position.

Can I build MightyMouse from source?

Yes. Clone the repository, run npm ci, then npm run dev. You need a Mac and a current Node LTS release. The README lists the build commands.

Read next