Dictate messages and notes
json
1 2 3 4 5 6 | |
Dictate messages and notes#
Don't feel like typing? Click the mic button, speak, and your words appear as text — ready to send or edit.
In the chat composer#
- Look for the microphone icon (🎤) in the message composer toolbar, between the emoji picker and the autocorrect toggle.
- Click once to start recording. A VU meter appears showing your audio level + elapsed time.
- Speak your message. Take your time — the recording captures everything until you stop it.
- Click the mic icon again (now showing a stop square) to finish.
- After a brief processing moment (transcription via ScaiEcho), your spoken words appear in the composer text field.
- Review and send — edit if needed, then press Enter/Send as usual.
The text inserts at the cursor position, so you can dictate into the middle of a message you're already composing.
In the note editor#
- Open any note in the editor.
- Click "Dictate" in the note toolbar (next to the voice-note button).
- Speak. The VU meter confirms your mic is active.
- Click again to stop.
- Your transcribed text inserts at the cursor position in the editor — Yjs broadcasts it to collaborators automatically if the note is shared.
Tips#
- Speak clearly — the transcription engine (ScaiEcho) works best with natural-paced speech, not rushed.
- One sentence or many — there's no time limit. Speak as long as you like; the entire recording is transcribed at once.
- Edit before sending — the composer is your review surface. Fix any transcription errors before pressing Send.
- Language — ScaiEcho auto-detects language. For best results, speak consistently in one language per recording.
What about the "Record voice" button in notes?#
That's a different feature — it records a full voice note, sends it through AI-powered transcription + organization, and produces a structured Markdown document. "Dictate" (the mic button described here) just inserts raw transcribed text at the cursor. Use whichever fits your workflow.
Where to go next#
- Talk to your AI — full voice conversations
- Todos from notes — voice-dictate todo items