Voice Input in Remote Mode
In remote control mode, the MIC key records with the keypad's microphone and sends the audio to a Whisper-compatible endpoint you configure. The transcript appears in an on-screen editor, and you press ACCEPT to submit it to the terminal.
An ASR service is required
Remote mode always records on the device, so the MIC key does nothing until you configure a speech-to-text service. The "prefer built-in ASR" setting applies to keyboard mode only.
Configure the speech-to-text service
Open the VibeKeys setup page in Chrome or Edge, connect to your device, and fill in the ASR (Speech Recognition) fields:
| Field | What to enter |
|---|---|
| URI | A Whisper-compatible transcriptions endpoint, e.g. https://api.groq.com/openai/v1/audio/transcriptions |
| API key | The API key for that service (sent as a bearer token) |
| Model | The model name, e.g. whisper-large-v3 |
Groq, OpenAI, and self-hosted whisper.cpp or faster-whisper servers all work.
Choose a MIC mode
The setup page offers two trigger styles for the MIC key:
- PTT (push to talk): hold
MICto record, release to send. - Toggle: tap
MICto start recording, tap again to stop and send.
Use it
- Press
MICand speak. The screen showsconnecting...in a yellow frame, then a green frame while it listens. - When recognition finishes, the transcript opens in an editor on the device screen.
- Move the cursor with the rotary knob to edit it.
- Press
ACCEPTto send the text to the terminal, orESCto discard it.