Using Dictation

Once voice is set up, press your hotkey and speak.

Starting a session

Press ⌘E (or your configured hotkey). A pill-shaped overlay appears at the bottom of the screen and moves through three states:

State What you see
Starting “Starting…” while the model loads
Listening A pulsing red dot while Verne captures your speech
Finishing “Finishing…” while the final result is processed

Stopping

How you stop depends on the mode you chose in Settings:

  • Toggle mode: press the hotkey again.
  • Hold mode: release the key.

The overlay disappears once the transcript is inserted.

Where text is inserted

Verne inserts the transcript at whatever has focus when you pressed the hotkey:

  • Terminal: text is sent to the active terminal session. The main use case: dictate a prompt directly to a running agent.
  • Monaco editor: text is inserted at the cursor position in the open file.

Post-processing

The post-processing you configured applies to every insertion. Spoken numbers become digits, developer terms are rewritten (e.g. “type script” becomes TypeScript), and any custom replacements you defined run in order before the text lands.

Next steps