Vidu S1 is a real-time interactive video generation model that lets users control digital characters and video content with voice instructions during generation. Built with TurboDiffusion and TurboServe, it reportedly produces 540p video at up to 42 FPS on regular consumer GPUs. The system supports custom images of real people, anime characters, and pets, along with selectable voice tones. The paper also claims infinite-length generation without blur, drift, or visual distortion, and reports the best results across all evaluated metrics while meeting real-time inference requirements. An online demo is available at Vidu Stream.
No heat snapshots are available in the last 24 hours.