Models
ByteDance's new-generation joint audio-video generation model, up to 30 seconds at a time and extendable twice; it understands the composition and camera language of reference videos, and supports professional editing like green screen and white model.