Reference to video
Attach up to nine images — a product, a face, an outfit, a location, a style frame — then type @ in the prompt to call them. Seedance 2.5 keeps every reference on-model for the whole take.
First and last frame
Upload a start frame and an optional end frame. Seedance 2.5 fills in the motion between them — a push-in, a reveal, a transition — so the clip lands exactly where your edit needs it.
Made for
From a 9:16 post to a 30-second spot — remix any example with its exact prompt.
Vertical clips for TikTok, Reels and Shorts — up to 30 seconds, with sound already in place.
Keep products and brand characters on-model from reference images — no shoot, no editor.
Previz shots with first and last frame control, and B-roll with consistent characters.
How it works
Open Create Video and choose Seedance 2.5 in the model picker.
Describe the shot, upload a start and end frame, or attach up to 9 images and call them with @.
Choose length (up to 30 s), format and quality — then generate and download MP4.
Specs and price
Longer and 1080p clips cost more — the generator shows the exact cost before you click Generate.
All in the same workspace — run one prompt on several models and keep the best take.
Seedance 2.5 is ByteDance's flagship AI video model. It turns text, an image or up to 9 reference images into one continuous shot of up to 30 seconds at up to 1080p, with sound generated together with the picture. On animx it is included in every paid plan, alongside Veo, Kling, Wan and other top models.
Start creating
Seedance 2.5 and every other top video and image model — one subscription, one workspace.