seedancetwo2.0: The video model that understands what you mean

seedancetwo2.0: Seedance Two accepts four modality inputs — image, video, audio, and text — for richer expression and more controllable generation. Reference Image to Video Precisely reproduce composition, character details, and visual style from a reference image to generate video. Reference Video to Video Replicate camera movements, complex motion rhythms, and creative effects from a reference video to generate new video content. Video Extension Smoothly extend and connect shots — generate continuous scenes guided by your prompts. Not just generation, but "keep shooting." Video Editing Swap characters, remove or add elements in existing videos with targeted edits.