Skip to main content
When paired with a custom AI speaker, Descript’s Regenerate feature can generate text-to-speech audio to fix a flubbed delivery, a misspeak, or a wrong spoken word in the original audio. If you’re looking to smooth an edit or fix pacing without changing what was said, see Use Regenerate to fix or smooth out any jumpy, abrupt, or awkward audio.
Before getting startedTo replace what was said, the original speaker must have a custom AI speaker created and authorized for use in Descript. If a custom AI speaker is not authorized and assigned Regenerate can’t be used to make script changes or alter the original audio.  Learn how to create a new custom AI speaker.

​
How to edit and replace original audio with Regenerate

  1. Add a speaker label and assign a custom AI speaker.
  2. Highlight the section of script you’d like to update.
  3. Click Regenerate in the selection toolbar, or press D.
  4. Type the new word/s, then click Regenerate to generate updated audio using the assigned AI speaker.
Descript will process the update and notify you once regeneration is complete. The new audio will appear in your timeline and stay synced to the updated script.

​
What happens if there’s video?

Regenerate will try to adjust the timing of your on-camera video to match the updated audio, but results may vary.

​
Requirements and limitations

  • A custom AI speaker is required. Regenerate does not support stock voices.
  • Regenerate only supports English audio
  • Up to 250 characters per regenerate action
Explore more Regenerate workflowsLearn how to smooth audio edits.