Tidy
Keeps your words. Takes out the ums, repeats and false starts, fixes the punctuation, and breaks a long dictation into short paragraphs.
Free, open-source dictation for Mac. Hold a key, speak, let go. Clean text appears wherever you are typing, and your voice never leaves your Mac.
Build 0.5 passed on CI, nice work everyone.
Are we still on for Friday? Anything left?
Before Friday we still need to update the changelog, bump the version and tag the release
I can do the changelog today
Still to do before Friday: update the changelog, bump the version, tag the release.
I'll do the changelog today.
Hi Sam,
Before Friday we still need to update the changelog, bump the version and tag the release.
I can do the changelog today.
Hi Sam,
We have three things left to do before Friday: update the changelog, bump the version and tag the release.
I can take the changelog today.
Friday release
Before Friday we still need to:
I can do the changelog today.
Before Friday:
Heard um so before friday we still need to uh update the changelog bump the version and and tag the release i can do the changelog today
A demo with example text. Nothing is recorded and no speech recognition runs on this page. A still from the demo, with example text.
Types wherever your cursor is
The film
Slack, Mail, Claude, your code editor. The text goes wherever your cursor is.
Speech becomes text on your Mac, with whisper.cpp on the GPU. Nothing is recorded or kept.
No account, no subscription, no trial. Nothing to sign up for, ever.
Every line is on GitHub. Read it, change it, or build the app yourself.
How it works
Choose sentence by sentence, by which keys you hold. Both can be changed, and dictation can be the fn key on its own.
Keeps your words. Takes out the ums, repeats and false starts, fixes the punctuation, and breaks a long dictation into short paragraphs.
Says what you meant, better, in a style you pick or one you write yourself. With text selected, it changes that text the way you ask.
Hold fn to talk, double-tap it to go hands-free. Utter keeps the emoji picker out of the way while it runs, and gives it back when you quit.
Writes for where you are
Utter knows the app in front, and in a browser the site. A chat message, an email, a code comment and a prompt each come out the way they should.
um so can we uh move the call to tomorrow like afternoon and can you send the deck before then
In a browser it reads the site too, so Gmail is an email and claude.ai is a prompt. Anywhere else you get plain, tidy prose.
can we move the call to tomorrow afternoon?
and could you send the deck before then
Could we move the call to tomorrow afternoon?
Could you also send the deck before then?
Can we move the call to tomorrow afternoon?
Please send the deck before then.
And everything around it
Fix a word it got wrong and Utter remembers it, with what it misheard. Next time “cloud” becomes Claude, but only when the sentence means it.
Select text, hold the rewrite key and say what to change. “Make this shorter.” “Fix the grammar.” The result replaces your selection.
Double-tap and talk for minutes with nothing held down. Tap once more to finish.
It reads the names in the app you are in, so it spells them the way the screen does.
Everything you said, kept on your Mac only. Search it, copy it again, or switch it off.
Its icon moves while it listens, so you always know. Fix the last word, add a word or copy your last dictation in one click.
A small pill says what Utter is doing and never takes your cursor. When it works, it simply disappears.
Inside the app
Privacy
Speech is turned into text on your Mac by an open model running on the GPU. Here is everything that ever leaves it.
Cleanup
A language model can tidy each dictation. Which one, and whether to use one at all, is up to you.
Or any OpenAI-compatible service. If cleanup ever fails, Utter types what it heard and tells you why.
Open source
Utter costs nothing and all of its source is public. Read every line, change what you like, or build it yourself.
Permission is hereby granted, free of charge, to any person obtaining a copy of this software…
Install
For Macs with Apple silicon on macOS 12 or later.
Open the DMG and drag Utter into Applications. It's signed and notarised, so it opens like any other app.
The microphone, so it can hear you, and Accessibility, so it can type for you.
The speech model downloads while you go. Then hold the key in any app and say something.
One line installs it, and brew upgrade keeps it current.
brew install --cask shantirmizi/tap/utter
The README lists the tools you need: Rust, Node and pnpm.
git clone https://github.com/ShanTirmizi/utter.git
cd utter && pnpm install
pnpm tauri build
Free and open source. No account to make and nothing to pay, ever.