Apple Watch Dictation to Mac and iPhone: Real Uses

When the wrist mic is the better mic

Watch dictation is a microphone placement decision. Your wrist ends up four to six inches from your mouth when you raise it, while a laptop mic sits two feet away behind a fan. That distance explains almost every case where the Watch produces a cleaner transcript.

• Loud room: a cafe, an open office, a kitchen with the extractor running. The close mic hears you and mostly ignores the room.
• On the move: walking to a meeting, waiting for a train, carrying a bag in the other hand. No phone unlock, no app hunt.
• Quiet speech: a library, a shared desk, a room with a sleeping kid. You can drop to a near whisper and still be understood.
• Docked Mac: the laptop is on a stand across the room, you are standing at the whiteboard, and reaching the keyboard means walking back.

Anything longer than a few paragraphs still belongs at the desk, where you can see the whole draft while you edit.

Set up Watch dictation once

You need the Auri app on your iPhone and Auri Voice on the Mac (macOS 15 or later). Pairing runs through a QR code, with no account sign-in involved.

1. Install Auri on your iPhone and add the Watch app from it.
2. Install Auri Voice on the Mac.
3. On the Mac: Auri Voice Settings > Apple Watch > Create Pairing. A QR code appears.
4. On the iPhone: Auri > Settings > Apple Watch Dictation, and scan the code.

That pairing covers the Mac destination only. Keyboard and Copy work as soon as the Watch app is installed. Sending to Keyboard does require Full Access for Auri Keyboard; the keyboard setup guide explains what that permission covers.

How a take works

The Watch app has one control: hold to talk. Speak the whole thought, release to send, and the screen steps through "Release to send" and "Transcribing…" to a confirmation that names the destination: "Sent to Auri Voice", "Ready in Auri Keyboard", or "Copied on iPhone". A carousel picks where the text goes:

• Mac: the text types itself at the cursor of the active Mac app. If nothing on screen accepts text, Auri copies it to the Mac clipboard and tells you to press Command-V.
• Keyboard: the text is ready in Auri Keyboard at the cursor the next time you open the keyboard on your iPhone. Requires Full Access.
• Copy: the text lands on the iPhone clipboard, ready to paste anywhere.

For the Mac destination, your Mac must be awake with Auri Voice running when you speak; there is no queue that delivers a take when you open the laptop later. Away from the Mac, choose Keyboard or Copy instead.

A take from the whiteboard, with a Slack thread focused on the docked Mac:

You said: "hey the retro is um moving to two, no, two thirty today, same room"

Auri typed: "Hey, the retro is moving to 2:30 today. Same room."

The cleanup pass dropped the filler and applied the spoken correction before the text landed. Focus works exactly as it does for shortcut dictation: the text goes where the Mac's cursor is, so click into the field before you raise your wrist.

Sending to the iPhone instead

The iPhone destinations cover the moments the Mac cannot: a walk, a wait in line, a hallway between meetings.

Keyboard suits replies. Dictate the answer while walking, open the conversation on the phone, switch to Auri Keyboard, and the text is waiting at the cursor. Copy suits notes: paste the take into Notes, your task app, or a draft when you sit down.

Two habits keep wrist captures usable later. Say one topic per take, and start each take with a two-word label such as "pricing page", so the paragraph explains itself when you paste it an hour later. If you collect several takes on one walk, paste each into a note as you go, because the clipboard only holds the newest one.

What happens to the audio

Watch-to-Mac audio is end-to-end encrypted in transit and transcribed on the Mac by a local engine. Auri will not route Watch audio through a cloud engine ("Watch dictation requires a local voice engine."), and the in-app retention line is short: audio is encrypted and deleted after 15 minutes. Transcripts stay in History on the Mac; on a shared machine, review Settings > Dictation > History.

Encryption protects the pipeline and does nothing about the person next to you. Passwords, one-time codes, salary figures, and client names deserve a keyboard or a closed door.

Loud rooms and quiet speech

In a noisy room, microphone distance does more for accuracy than engine choice. A laptop mic across the table records you and the espresso machine at similar volume, and no speech model reliably separates them. The wrist mic changes that ratio before recognition starts.

• Raise the wrist and speak across the mic rather than straight into it. Breath hitting the mic head-on causes pops that come out as dropped syllables.
• Turn your body away from the loudest source, even by 45 degrees.
• Keep the cuff and sleeve clear of the microphone opening.
• Speak in complete sentences so context helps the model resolve blurred words.

Quiet speech works for the same reason the loud room does: the mic is close, so a near whisper still carries enough signal. Keep normal sentence rhythm, finish words fully, and say the whole thought in one take rather than three whispered fragments. If the wind is loud enough that you would raise your voice on a phone call, expect to fix a word or two afterward. The speaking guide covers pacing and phrasing in more depth.

Limits worth planning around

Watch dictation stays a short-form tool. Battery, wind, wrist position, and a sleeping Mac cause more friction than recognition quality does.

FAQ