Seedance 2.0 Mini image-to-video from ByteDance starts the clip on a still you supply and animates it, with sound in the same pass. Clips run 4 to 15 seconds at 480p or 720p, and an optional closing still makes the shot arrive on a set frame.
Workbench
These examples were written for this endpoint. Replace the prompt and sample media URLs with your own accessible inputs, then check the current quote and model availability before submitting.
- Output5s
bioluminescent-tide-glance `image_url` is the opening frame: a woman in white silk stands at the edge of a night beach where the waves glow electric blue, the Milky Way overhead and palms at the edge of the frame. The clip plays the prompt in order: she walks slowly into the glowing shallows, each step splashing blue light up around her ankles, then turns and looks back over her shoulder at the camera with a slow smile while the camera stays low behind her. `aspect_ratio` `9:16` keeps the vertical frame of the opening image, and `generate_audio` is `true`, so the clip comes with its own sound.
- 720p
- 9:16
- audio
- Output4s
latte-then-robe One tilt, nothing else. The opening frame is a tabletop; the prompt lifts the cup away, moves the camera up and stops. Keeping an image-to-video brief to a single camera move is what makes the smallest tier dependable — every extra event you add is another thing that can arrive in the wrong order. `duration_seconds` sits at `4`, the shortest this endpoint accepts, which is all one tilt needs.
- 720p
- 3:4
- 4s
- Output4s
shelf-restock One `image_url`, one small action staged down to which side the hand enters from and which way the cans face: "enters from the right", "facing each label forward", locked camera. The output plays it exactly — a glove fills the gap can by can, labels turned out, then withdraws, shelves above and below untouched. `duration_seconds` sits at `4`, all one gesture needs.
- 720p
- 16:9
- audio
- Output10s
season-turn Two stills bookend a season change: leaves turn, light drops, the hedge thins. Naming what changes, not what happens, lets the model treat all ten seconds as one shift. `last_image_url` supplies the autumn end point. The output keeps the road and hedge locked between the two photos while colour and light carry the passage of time. `duration_seconds` is `10`.
- 720p
- 16:9
- audio
This example starts at $0.025 per output second. The Playground updates the estimate when you change the settings.