studio & library/Avatars

Avatars

Consistent on-camera characters for lip-synced video. Currently in beta.

VIDEO WALKTHROUGH · 4 min
Creating and using an avatar

Building an avatar from photos and from a prompt, then rendering a scene with the Narrator renderer.

Recording in production — the written steps below cover the same ground.

Three ways to get one

Use a default
A curated set of ready-made presenters.
Upload photos
Your own images, to keep a real face consistent across videos.
Generate from a prompt
Describe the presenter and generate the face.

Scenarios

An avatar can be placed into different scenarios — settings and framings — and generated from an existing avatar so the same character appears in a new context without losing their likeness.

Using an avatar in a video

Pick a host during channel setup, or assign one to any scene using the Narrator renderer at the Direction gate. The avatar’s face plus the scene’s narration produce a lip-synced talking-head clip.

Talking-head is the most expensive kind of generation in the system, and it is a paid capability — the free tier has none. Budget for it when planning a presenter-led channel.

Organising

Search by name and star the ones you use most; the Favorites filter narrows to those.

Related
PreviousText to SpeechNextChannel workspace
Didn’t find your answer?
Ask us directly and we’ll walk you through it — your question usually becomes the next page here.
Contact us