Seedance 2.5 Video Extension continues an existing timeline shot after the selected boundary. Kepler renders up to 30 seconds of the video immediately before the Extend region and uses it as motion, scene, character, and audio context.
The source context must be at least 1.8 seconds long. Generated duration can be set from 4 to 30 seconds. Resolution offers Match Canvas, which chooses the nearest supported 480p, 720p, or 1080p output for the project canvas. The model follows the reference video aspect ratio, so Aspect Ratio is not shown. Seedance extends forward only, so the unsupported Before direction is hidden.
With Native Audio enabled, Seedance can continue ambience, sound effects, speech, and music together with the picture. For lip-synced dialogue, put the spoken line in quotation marks and identify the speaker clearly. Lip movement, timing, identity, and exact continuity can still vary between generations.
Write one chronological prompt describing what should happen next, including subject movement, camera behavior, sound, and dialogue. Advanced settings expose output Bitrate and an optional Seed. Negative Prompt, Prompt Strength, Elements, and separate shot controls are not accepted by this endpoint and are not shown.
Provider billing includes both the rendered video context and the generated continuation. Kepler shows the complete credit estimate before generation. The returned continuation is placed on a new generated track in the Extend region; the original timeline media is never overwritten.