Editor Dictation

Dictate directly into any open editor — source files, markdown, commit messages, or any text input.

How It Works

1. Place your cursor where you want text to appear
2. Press Cmd+Shift+D (macOS) or Ctrl+Shift+D (Windows)
3. Speak — a light "tink" tone confirms recording started
4. Press the same key combo to stop — text is inserted at all active cursors

Multi-Cursor Support

If you have multiple cursors active (e.g., Cmd+D to select multiple occurrences), dictated text is inserted at every cursor position simultaneously.

Insert Mode

By default (final), text is inserted only after you stop recording. Change to streaming to see text appear as you speak:

{
  "l33t-speak.insertMode": "streaming"
}

Mode	Behavior
final	Inserts all recognized text when you stop recording
streaming	Inserts interim results in real-time as you speak

Cancel

Press Escape to cancel recording — nothing is inserted and the audio buffer is discarded. You'll hear a low "basso" tone confirming the cancel.