Push-to-talk is the default. Optimized for short bursts wherever your cursor is.
How it works
- Click into a text field.
- Hold
Fn. - Speak.
- Release.
The HUD shows a live waveform and timer while you hold the key. On release, the transcript pastes into the active field automatically.
Side-by-side: a Slack compose box with cursor, and the Whiskers HUD showing a waveform mid-recording
Where it works
Anywhere there's a text input that accepts paste. No per-app setup — Notes, Slack, Mail, browser, IDE, Terminal, all work the same.
If a rare app blocks programmatic paste, the transcript stays on your clipboard and you can Cmd+V it manually.
Limits
| Limit | Value | Behavior |
|---|---|---|
| Minimum | 300ms | Quick taps are ignored — prevents accidents |
| Maximum | 10 minutes | Auto-stops and processes |
For recordings over 10 minutes, use File Transcription on a pre-recorded file.
Cancel
Press ESC at any time. Audio is discarded; nothing is saved to history.
When to use it
- Short sentences, quick thoughts.
- Slack, chat, comments.
- Anytime you want explicit start/stop.
For paragraphs and long-form, hands-free mode is more comfortable.