Avatars let you add an on-screen presenter to your videos without recording on camera. Depending on your workflow, you can generate a voice with text-to-speech, record your own voice, or pair avatars with imported audio.

This article covers three different ways to create and use avatars in Descript, with step-by-step instructions for each workflow:

• Record your voice and assign it to an avatar
• Create an avatar video from scratch with text-to-speech audio
• Add an avatar to existing audio or media

For an overview of what avatars are and how they work, see our Avatars Overview guide.

Record in Descript first, then assign an avatar

Workflow: Avatar + Your recorded voice

If you want to appear in your video using your own voice—without being on camera—you can record your voice directly in Descript and pair it with an avatar.

1. Open a new project and click Record at the bottom of the window.
2. Record your content directly into your Descript composition. Need help recording? View our recording guide.
3. After recording, click the Add speaker button at the top of the script to apply a speaker label. You’ll have three options:
   • Choose an existing custom AI speaker.
   • Use one of Descript’s stock AI speakers.
   • Create a new custom AI speaker.
   
4. Make sure the label is applied to your script before assigning an avatar. Then, click the speaker label again and select Assign avatar to [your speaker name].
5. Create a custom avatar with a text prompt, select a stock avatar from the gallery, or upload a headshot. Learn more about creating avatars in Descript.
6. Click Generate avatar to animate it with your recorded voice. This should be a final step before finishing your composition — once you generate the avatar, any additional edits will require re-generating and an additional debit of your avatar minutes.

Create a video from scratch with TTS audio

Workflow: Avatar + Text-to-speech voice

Write a script, generate text-to-speech audio, then select a stock voice or existing custom AI speaker, and assign an avatar to present your content.

1. Open a new video project/composition and click Create with AI speaker. 
2. Select a stock voice or existing custom AI speaker. Click Choose speaker to confirm.
3. Select a stock avatar from the gallery, or create a custom avatar. Click the Assign avatar button to confirm your selection. Learn more about creating avatars in Descript.
4. In Write mode, draft or edit your script. When you're done, click Done writing.
5. Click Generate avatar to animate it with your text-to-speech voice. This should be a final step before finishing your composition — once you generate the avatar, any additional edits will require re-generating and an additional debit of your avatar minutes.

Add an avatar to existing audio or imported media

Workflow: Avatar + Imported voice track

Working with recorded audio—like a podcast, voiceover, or screen recording? You can pair it with an avatar to add a visual presenter.

1. Import your media into your Descript project and add it to your script. Learn how to import and transcribe your content.
2. If you haven’t already, add a speaker label to your script. Click the speaker label and choose Assign avatar.
3. Select a stock avatar from the gallery, or create a custom avatar. Click the Assign avatar button to confirm your selection. Learn more about creating avatars in Descript.
4. Click the Assign avatar button to confirm your selection.
5. Click Generate avatar to animate it with your recorded voice. This should be a final step before finishing your composition — once you generate the avatar, any additional edits will require re-generating and an additional debit of your AI credits.

Not seeing your avatar? Review the layer order.

Avatar visibility depends on layer order and visibility settings. Check that your avatar layer is turned on and sitting above other media in the Properties panel.