Your dictionary
Friends’ and colleagues’ names, your pets, street and shop names — add them once and MewFlow will spell them right.
Hold a key and speak. Let go — and the text, with punctuation and paragraphs, is already where your cursor is. In any app.
No sign-up, nothing to install — right here in the browser, up to 20 seconds. Speak any way you like: with “um”s, repeats and corrections on the fly.
um hi so listen let's do tomorrow at three no at four and I'll send the meeting link
Nothing is stored: Groq recognizes the audio, Claude by Anthropic tidies the text, and that’s it.
Start recordingStop recordingPress the button and speakListening… Press again when you’re done.Processing…Done! Try again?Too short — say at least a few words.The browser didn’t allow the microphone. Allow it in the site settings and try again.Couldn’t turn on the microphone. Check that it’s connected.This browser can’t record audio. Try Chrome, Safari or Firefox.That’s enough tries for today 🙂 Install MewFlow — 7 days free.Too many people right now. Try again in a few minutes.That didn’t work. Please try again.We didn’t hear anything — speak closer to the microphone.CopyCopied ✓No commands to learn, no saying “comma” or “new line”. Talk the way you’d talk to a person — MewFlow adds the punctuation, drops the filler and splits the text into paragraphs.
You can pick another key. A quick tap turns on hands-free mode: talk as long as you like, then tap again when you’re done.
With pauses, “um”s and corrections on the fly. Ukrainian, Russian and English terms all mixed together — it gets it.
Usually within a second. Esc cancels without pasting anything.
About 50 languages for speech, plus mixed speech with English terms. The app itself speaks 20 languages. Pick a language and see what happens to raw speech.
«е-е слухай я скину звіт до п’ятниці ні до четверга і ще треба перевірити pull request»
Слухай, я скину звіт до четверга. І ще треба перевірити pull request.
We measured it on real dictations: recognition takes a fraction of a second, cleanup about the same again. Speaking is three times faster than typing — and now that’s not just for short messages.
Friends’ and colleagues’ names, your pets, street and shop names — add them once and MewFlow will spell them right.
Dictation audio and text aren’t stored on the server. Your history stays on your computer.
Say “first, second” and you get a list. MewFlow splits a long train of thought into paragraphs.
“At three, no, at four” — only “at four” ends up in the text.
Got an NVIDIA card? Speech is recognized right on your computer — fast, and the audio never leaves it.
From Ukrainian and Polish to Japanese and Hindi. The app speaks your language.
A voice keyboard for iPhone and Android is in the works. One account and one subscription on every device.
Try it free for a week, no card needed. Like it? It’s 199 UAH a month. Cancel anytime in one click.
Your dictation history lives only on your computer, and you can turn it off or clear it. The server processes the audio and text to return the result and keeps none of it — only usage counters.
In any text field on Windows and Mac: messengers, email, the browser, documents, code editors, spreadsheets. The text goes wherever the cursor is.
About 50 languages for speech, including mixed speech such as Ukrainian or Russian with English terms. You can pick the language or let MewFlow detect it. The app interface is available in 20 languages.
On a Windows PC with an NVIDIA graphics card recognition can run right on the computer. Tidying the text — punctuation, paragraphs, corrections — needs the internet. On a Mac everything works over the internet.
The Mac version is out — for Apple Silicon (M1–M4) and Intel. iPhone and Android are in the works; they’ll use the same account and subscription.
Yes: in “Try it right now” say anything for up to 20 seconds and see what MewFlow would paste. No sign-up. →
In your account, choose “Cancel subscription”. There will be no new charges, and you keep access until the end of the paid month.