Support

Press ⌘ ⇧ D from any application — a word processor, an email client, a browser. A small panel appears on top of your windows while you speak. Press the shortcut again to stop: the text is placed on your clipboard, and all that is left is to paste it with ⌘ V.

You can change this shortcut in Settings › Dictation.

On your Mac, in the space macOS reserves for the app. They are never sent anywhere. You can delete a single session, or every recording, from Settings › Data.

The full details are on the privacy page.

It depends on the engine selected in Settings › Transcription. Whisper, the default engine, offers 35 languages. Apple's engine offers only those your device can recognise, and requires macOS 26.

The exact number is shown in Settings next to each engine. It is counted from the list actually offered — never announced in advance.

The engine you choose is the engine that transcribes — the app never switches behind your back. It is the list of languages that follows the engine, and it never offers one the engine could not handle.

First check the microphone permission: System Settings › Privacy & Security › Microphone, and make sure ExvoxAI is listed and enabled.

Then check that the right microphone is selected in the macOS sound settings — a connected headset often takes over without you noticing.

Yes. Drop the file into the window, or use Import a file. Common audio and video formats are accepted. The app separates the different speakers and timestamps each passage.

Dropping an image or a PDF opens the document tools, which extract the text from it.

No. Transcription works offline. A connection is only needed in two cases: downloading an additional model, or turning on an online AI service to summarise text — which is optional and off by default.

Open Settings › Text and add the word to your correction dictionary. The spelling you set is applied to every transcript from then on — useful for people, places, or the vocabulary of your trade.

For a one-off fix, simply double-click the line in the transcript.

Because the sound is coming out of your Mac's built-in speakers. Plug in headphones and recording starts.

This is not a comfort setting. On speakers, your microphone hears the whole meeting back: its track then transcribes everyone else's words verbatim, and since everything from the microphone is labelled “You”, sentences you never said would be attributed to you. No sensitivity setting fixes that — hence the refusal.

Headphones, AirPods or a plugged-in speaker: everything else is allowed. The app only blocks on a certainty, never on a doubt.

They describe where the speech comes from, not what the app does with the text.

Dictation — you speak alone.
Meeting — everyone is in the room: the microphone is enough, and voices are separated.
Video call — Teams, Zoom, Meet: the app records your Mac's audio and your microphone on two separate tracks, so you and the others never blend together.

Install ExvoxAI on both devices, put them on the same network, then pair them from Settings › Remote microphone on the Mac.

In the iPhone's dictation screen, pick your Mac as the destination: what you say is written on the computer. Useful when the Mac is out of reach, or when you want to move away from it while speaking.

Nothing goes through the internet: the link stays on your local network.

ExvoxAI is bought once, with no subscription, and the purchase covers both the Mac and the iPhone. Download it on either one: it is already paid for.