Skip to main content
You hold a shortcut, speak, and release. Vowen records your voice, turns it into text with your speech model, cleans the text up, and inserts it where your cursor is. Everything below happens in a second or two. The recording indicator pill with a live waveform while you speak

Recording modes

You can change both shortcuts, switch a shortcut from Hold to Tap, add mouse buttons, and double-tap the push-to-talk key to lock a recording on. See Shortcuts and Shortcut patterns. A hands-free recording also stops when you click the stop button on the recording indicator.

Stop hands-free with the dictation shortcut

Any dictation shortcut stops a hands-free recording, however it was started. That includes the push-to-talk shortcut, its alternatives, tone shortcuts and mouse buttons.

Cancel a recording

To discard a recording instead of transcribing it:
  • Press Cancel Recording, which is Esc by default. The key press doesn’t reach the app you’re typing in.
  • Or press and hold the stop button on the indicator for half a second. A ring fills as you hold. Releasing early stops the recording normally.
A Transcription Cancelled notice appears with an Undo button for about five seconds. Click Undo to continue the same recording. If you don’t, the audio is discarded and nothing is transcribed.

The pipeline

1

Record

Your microphone records while the shortcut is held or while hands-free is on. The recording indicator shows that Vowen is listening.
2

Transcribe

The audio goes to your speech model. A local model runs on your computer. A cloud model sends the audio to that provider with your API key. Models that stream can show the text as you speak with real-time preview.
3

Clean up

Vowen corrects dictionary misspellings, applies your threads, and removes filler words such as “uh” and “um” when Settings › General › Remove filler words is on. Spoken voice shortcuts are checked here too.
4

AI enhancement (optional)

If you turned on AI enhancement, the text goes to your AI provider for grammar, punctuation and formatting. Threads set to run after AI are applied next.
5

Fit to the cursor

Smart spacing and capitalization adjusts spaces and the first letter to match the text around your cursor.
6

Insert

Vowen pastes the text into the focused field, or types it with direct insertion. See Text insertion.
The indicator in its processing state after you release the shortcut Every dictation is also saved to Home › History unless you turn off Settings › General › Save transcriptions to history. See History.

Smart spacing and capitalization

Right before inserting, Vowen reads the characters on either side of your cursor and fits the dictation into them. It adds a missing space before or after the text, and lowercases the first word when you are in the middle of a sentence. When it can’t read the surrounding text reliably, the dictation is inserted as spoken. Turn it on or off in Settings › General › Smart Spacing and Capitalization. It is off by default. The setting is per device and is not synced. The Smart Spacing and Capitalization row in Settings › General The surrounding text is read on your computer and is not sent anywhere. Text insertion has the full rules and examples.

Silence

Local Whisper models skip silence with voice activity detection, which also stops the model from inventing text such as “Thank you” on an empty recording. Parakeet recordings shorter than 300 ms are discarded. If a cloud model still produces text from background noise, turn on Settings › Experimental › Enhanced silence detection and adjust its sensitivity.

Sound effects

Vowen plays a short sound when recording starts and stops. Turn it off, or change the volume, in Settings › Audio › Sound Effects and Sound Effect Volume. See Microphones.