All guides

Voice input

Dictate into the prompt bar, the chat panel or a note, with speech recognition that runs on your PC.

Set up the microphone

Voice input needs a one-time download of about 0.5 GB. After that it works without internet.

  1. Open Settings, then Models, and find Voice input. Download it and leave it on.
  2. Open the Voice input row. Under Microphone, choose Default, Communications or a specific microphone.
  3. Check Languages. Voice input understands any of Your languages that it supports. If it supports none of them, it shows “None of yours” and can’t run.
  4. If Windows blocks the microphone, allow desktop apps to use it in Windows microphone privacy settings.
Link to this section

Speak into a bar, chat or note

Click the microphone to start and Stop to finish. Words appear as you speak. Wait for the last words to settle before you press Enter.

Turn on “Hold to record” in the Voice input row to listen only while you hold the microphone button. It works with the mouse button only. Right-click the microphone to change the microphone or “Hold to record”.

Typing stops listening. In the prompt bar, Escape closes the bar and drops the recording. Enter sends the words recognized so far and drops any still being recognized. In a note, words go in at the cursor.

Link to this section

Start listening automatically

In the Voice input row, turn on Listen when chat opens or Listen when a note opens to start listening as soon as they open.

In a pill, add an Input action set to Type it and turn on Listen right away. The pill’s prompt bar then opens already listening.

Link to this section

Supported languages

Voice input understands Swedish, English, Norwegian Bokmål, Danish, Finnish, German, French, Spanish, Italian, Portuguese, Dutch, Polish, Estonian, Czech, Romanian, Russian and Ukrainian.

Your languages can list more than these. With several supported languages selected, Voice input picks the language you’re speaking.

Link to this section

What stays on your PC

Speech is recognized on your PC, and your voice isn’t uploaded. The words go into the field you’re using and then behave like typed text: a note is saved on your PC, and a question to an online AI model sends those words to that provider.

Link to this section

If voice input won’t start

  • No microphone: connect or turn on a microphone, then choose it under Microphone.
  • Microphone blocked in Windows: click the microphone again to open Windows microphone privacy settings and allow desktop apps.
  • Not enough memory: close other apps and try again. Voice input needs about 1 GB of free memory.
Link to this section