Seedance 2.5 Motion Control creates a new video from two references. Select a still image for the subject and a 4-30 second timeline region for movement. Kepler renders that region as a video reference. Seedance uses the video to guide movement, timing, framing, and camera work while the image guides the subject's appearance. The result is placed on a new generated track; source media is not overwritten.

In the Motion Control window, choose Seedance 2.5 and pick the still image. Duration also defines the video reference region. Resolution offers Match Canvas plus 480p, 720p, and 1080p. Native Audio generates synchronized sound with the new video; it does not copy the original soundtrack. Advanced settings contain Bitrate and an optional Seed. Seedance does not accept Orientation, Original Sound, Adapt Motion, Enhance Identity, or Negative Prompt controls, so those are hidden for this model.

Describe what should move and which traits of the image should remain. Kepler identifies the motion video as @Video1 and the still image as @Image1 in the request. For dialogue, name the speaker and put spoken words in quotation marks; Seedance can generate lip-synced speech when Native Audio is enabled.

The credit estimate includes both the rendered video reference and the generated output. Check the visible estimate before starting a generation. A clean, well-lit motion reference and a subject with a comparable body structure usually give the model clearer guidance.