Talk instead of typing. Tap the microphone in the brain dump, say everything on your mind, and read back the tasks it found. This needs the iPhone app.
How it works
- Open the brain dump: the + button, or ⌘J on a keyboard.
- Tap 🎙 Speak. The first time, iOS asks for the microphone and for speech recognition.
- Talk. The words appear in the box as you go, one thought per line.
- Tap Stop, read what came back, edit anything that came out wrong.
- Tap Capture. From here it is an ordinary capture: AI proposes tasks with dates and priorities, and you approve each one.
Nothing is added until you have read it. A misheard word costs you an edit, never a wrong task in your list.
Your voice stays on your iPhone
Transcription runs on the device. No audio is uploaded, none reaches us, and no recording is kept. That also means it works with no signal at all.
The tasks themselves are a normal capture, so the text you approve is filled in by AI on the server, exactly as a typed capture is.
One language at a time
The microphone listens in one language per session. Pick it with the buttons under the box; the choice is remembered and follows you to your other devices.
Two languages in one sentence will not both come out right. Say the Korean parts in a Korean session and the English parts in an English one.
If the microphone is not there
- No button at all: this iPhone has no on-device speech model for any language we can offer. Typing still captures everything.
- "Getting ready": iOS is fetching the language model. It happens once.
- Access is off: turn the microphone and speech recognition back on in iOS Settings, under nowdo.
Good to know
- A session stops itself after three minutes, and keeps every word.
- Voice capture is free on every plan. It costs nothing to run, because your phone does the work.
- Prefer to speak with the app closed? That is Siri capture, which takes one task at a time.