Hold a key. Talk. Keep your hands where they were.
Hold the right ⌥, say what you mean, let go. The words
appear in whatever app you were already in — the reply, the brief, the
commit message, the field you were halfway through filling in. Hold the
right ⌘ instead and the same sentence comes out
translated.
Requires macOS 15 or later on a Mac with Apple silicon · signed and notarised by Apple
On device
the model runs on your Mac
Two keys
dictate, or dictate translated
Any app
wherever the cursor already is
Click to enlarge
The whole interaction
There is no window to go to
Dictation tools usually ask you to leave what you are doing, speak into
their window, then copy the result back. This one is a key you hold.
Press and speak
Two keys, two jobs
Right ⌥ — say it as you said it
Hold it and talk. Let go and the text is inserted at the cursor, in the app that was in front. Nothing to open, nothing to paste, nothing to switch back from.
Right ⌘ — say it in one language, send it in another
The same hold-and-talk, with a translation step on the way out. Speak Polish, an English sentence arrives. Translation is a separate key rather than something the app guesses at, which is the point — see below for why.
Escape cancels it
Started a sentence you did not want? Press Escape while still holding, and nothing is transcribed and nothing is inserted.
The unglamorous half
Getting the text in without breaking anything
Your clipboard is put back
Insertion goes through a paste, so the clipboard is saved first and restored afterwards. Whatever you had copied is still there when the sentence lands.
A fallback for apps that refuse a paste
Some apps will not take a synthetic paste. Those get the text typed in character by character instead, so the answer to "it does not work in this one app" is not a shrug.
Your first word is not eaten
A microphone needs a few hundred milliseconds before it delivers usable audio, and by then you are already talking. The microphone is kept warm between recordings and half a second of audio is held back, so the beginning of the sentence is there.
Decisions you can feel
Built against the ways this goes wrong
Every item here exists because it is a common, specific complaint about
dictation tools — not because it looked good on a feature list.
The shortcut does not quietly stop firing
The most common failure in this category is the hotkey dying mid-session and never coming back until a restart. The key is watched at a level that can be re-armed, and it is re-armed when the system switches it off.
Your language is set, not guessed
On a short sentence, automatic language detection often picks wrong — and a wrong guess does not produce a typo, it silently produces a translation. So the dictation language is something you choose, and translating is the other key.
Silence is not sent to the model
Speech models invent sentences when handed silence — subtitle credits, thank-yous, whole polite closings that you never said. Audio with nothing in it is not passed on, so it has nothing to invent from.
Nothing waits forever without saying so
Loading a speech model onto Apple's Neural Engine can hang with no error and no timeout, at nought per cent processor, indefinitely. Loading has a deadline; when it passes, the app loads a different way and remembers to do that next time. An app that goes quiet for an unbounded time is indistinguishable from a broken one.
It lives in the menu bar
No dock icon to alt-tab past, no window in your way. A small indicator while it is listening, and otherwise nothing.
Signed and notarised by Apple
Distributed directly, with a Developer ID signature and Apple's notarisation, so macOS lets it open instead of refusing it the way it refuses an unsigned download.
Where the audio goes
Nowhere.
Speech is turned into text by a model that runs on your own Mac, using
the machine's own hardware. Your voice is not uploaded to us, and it is
not sent to any speech service.
Audio is not storedIt exists for as long as it takes to turn it into text, and then it is gone. There is no recordings folder and no history to leak.
One download, at the startThe speech model is fetched once when you first set the app up, and after that dictation works with the network off — no sentence you speak is ever sent anywhere. The only other time the app goes online is to activate your licence key, once.
Client work stays client workThis is the reason the model is local rather than an API call. If you are under an agreement that forbids sending a client's words to a third party, a cloud transcriber breaks it and this does not.
What the app does contact us about, once you have bought it, is your
licence key — see the Privacy Policy.
Straight answer
What is not finished
This is an early build.
Rather than describe it as though it were finished, here is where it
actually stands.
Translation needs a recent macOSThe second key works, and translates Polish to English out of the box. It does not go through the speech model — the words are transcribed first and then handed to Apple's own translator, which is part of the system. That translator is the catch: it arrived in macOS 26, so on anything older the app dictates but the second key does not translate.
The first run waits on a downloadDictation does not work until the speech model is on your Mac — 460 MB, fetched once. On a slow connection that is a wait with nothing to do, and the app has no way to shorten it. Everything after that first fetch is local and needs no network at all.
The orange dot stays onmacOS lights it whenever an app holds the microphone, and this one holds it all day deliberately: opening the microphone takes a few hundred milliseconds, which is exactly the time your first word would be lost in. So the indicator is lit while nothing is being recorded, and there is no honest way to have both a lit-only-while-recording dot and a first word that survives. Audio lives in a half-second buffer that is continuously overwritten, and nothing is written to disk or sent anywhere unless you are holding a key.
You pick the one you speak rather than letting the app guess, and it
stays picked. Every one of these is transcribed by the model running
on your own machine — none of them is sent anywhere.
🇧🇬Български
🇭🇷Hrvatski
🇨🇿Čeština
🇩🇰Dansk
🇳🇱Nederlands
🇬🇧English
🇪🇪Eesti
🇫🇮Suomi
🇫🇷Français
🇩🇪Deutsch
🇬🇷Ελληνικά
🇭🇺Magyar
🇮🇹Italiano
🇱🇻Latviešu
🇱🇹Lietuvių
🇲🇹Malti
🇵🇱Polski
🇵🇹Português
🇷🇴Română
🇷🇺Русский
🇸🇰Slovenčina
🇸🇮Slovenščina
🇪🇸Español
🇸🇪Svenska
🇺🇦Українська
Before you count on it
What it needs
A Mac with Apple silicon
The speech model runs on the Neural Engine in Apple's own chips, so an Intel Mac is not supported.
macOS 15 to dictate, macOS 26 to translate
Dictation runs on macOS 15 and later. Translation is Apple's own, built into the system, and that piece only exists from macOS 26 — below it the second key has nothing to call.
Microphone access
Granted once, in System Settings, the first time you dictate.
Accessibility and Input Monitoring
Also granted once. They are what let the app notice the key you are holding and put text into another application — the two things it is for.
Room for the model, and one download
460 MB on disk, fetched once during setup.
Bought and updated here, not from the App Store
Holding a key down anywhere on your Mac, and typing the result into whatever app you are in, is exactly what the store's sandbox forbids. So this one comes straight from us, signed with a Developer ID and notarised by Apple, and it updates itself from here. Nothing about that is temporary.
One price, forever
Dictate everywhere on your Mac.
Seven days free, then one payment — no subscription, and the version you buy keeps working. Fourteen days after that to change your mind, with no reason needed. The key arrives by email in a couple of minutes, and there is no account to make.
