OmniHuman 2.0 AI Video Generator

OmniHuman 2.0 is BytePlus’s audio‑driven avatar video model upgraded for up to 5‑minute single shots. Taking one reference image and audio, it interprets rhythm, emotion and semantics to generate lip sync, expressions and gestures. Text prompts control camera movement and acting style. Multi‑character scenes are available for narration, music MV and virtual IP content.