Mirage’s Avatar X claims a digital twin from 10 seconds of video

Scobleizer · x · 2026-07-29

Mirage’s Avatar X claims it can create a convincing digital twin from just 10 seconds of video, with strong identity preservation and facial micro-expressions.

The post says the model handles not only lip sync but also non-verbal cues like laughing, crying, and yawning, with expressions driven across the face and body by audio. It is framed as a clear step beyond avatars that only animate the mouth while leaving the eyes still.

Related event: Mirage launches Avatar X for high-fidelity AI digital twins(12 posts)→

Original post →

More from Multimodal

Multimodal channel →