Skip to content

Set the default Whisper model

Item Current behavior
Default large-v3-turbo
Valid values An installed Whisper model from the current in-app catalog.
Existing sessions Keep the model recorded when the session was created.
New sessions Use the selected default for import, recording, and voice-input transcription.
Storage synthezia-settings in webview local storage.
Privacy Whisper runs locally; downloading a missing model is a separate network action.
Restart None after selecting an installed model.
  1. Open Models with Cmd+M.
  2. In Transcription, select an installed model. Select a catalog model first to download it.
  3. Start a new action to use the selection.

If the preferred default is missing when an action starts, Synthezia opens a model-requirement dialog instead of transcribing without a model.