Interaction Modes

⚠️ Note:

• Seedance 2.0 provides entry points for Single Image, Start / End Frame, and Omni Reference (with Smart Multi-frame and Subject Reference enabled). If you need multimodal input (image, video, audio, text) in one generation, use Omni Reference.
• The current interaction pattern is to specify how each image / video / audio is used via “@asset name”, for example: @Image1 as the first frame, @Video1 for camera motion reference, and @Audio1 for background music.