Omnihuman API Documentation

Generates hyper-realistic full-body videos from a single image and audio/video input using advanced diffusion transformers, supporting diverse styles, natural motion, and HD outputs up to 1024×1024 at 30fps.