Voiceover

Voiceover generates AI voiceover audio from text so narration, placeholders, dialogue, and production notes can move through the edit faster. It is useful when the timeline needs a voice track before recording, or when a generated read is good enough for the current version.

Generated voiceover is stored as new audio media. It does not replace existing audio unless you choose to edit it into the timeline yourself.

A typical workflow:

- Open Voiceover.
- Write or paste the script.
- Choose the voice model and available voice settings.
- Add reference audio only when the selected model requires or supports it.
- Review the credit cost before generation starts.
- Place the generated audio in the timeline and trim it to the edit.

Voiceover quality depends on the script and settings. Write punctuation the way the voice should breathe, keep lines readable, and use direction fields for tone or delivery when available. Review generated speech for pronunciation, pacing, and loudness before using it in a final export.