Add high-quality voiceover to your videos without recording. Give a Speaker a stock voice and Descript generates natural-sounding audio for anything you write under that label.
Use a stock voice
- Click Add speaker above a paragraph that doesn't have a Speaker label yet, or place your cursor anywhere in the script and press
@. - Select Browse stock AI speakers.Â
- The picker opens on Custom speakers (any voices you've created) followed by Recommended (a short curated list of stock voices.)
- Click the play button next to any voice to hear a sample.
- Click a voice to select it, then click the Choose speaker button to confirm.
Descript will automatically begin generating text-to-speech audio for the script that follows the Speaker label.
Search and filter the voice catalog
Enter any search term to peruse the voice catalog. Click the See more button to open the full library of available voice options.
- Search: enter any term here to target the type of voice you’re looking for
- Filter: narrow by Gender, Age, or Language.
- See more: scroll to the bottom of the Recommended list and click See more to search the full catalog.
Language compatibility and non-English voices
Descript offers stock voices in a variety of languages. To get the best results, use a voice whose tagged language matches the language of your script.
Some voices are designed for a specific language
Audio may still generate if the languages don't match, but it may sound less natural or lose the intended tone. See supported text-to-speech languages
Legacy stock voices
Some older voices have been moved to a Legacy category as Descript's AI models improve. These voices still work, but may sound less natural.
To view available legacy speakers:
- Open the voice picker.
- Click the filter icon.
- Change Legacy speakers from Hide to Show.
Replace recorded audio with a stock voice
Stock voices can also replace your original audio. Descript generates the new voiceover from your script text and keeps your original video in the composition. For that workflow, see Convert recorded audio to AI speech.