Seedance 2.0 Fast Reference-to-Video

Generates a video from a mix of reference images, videos, and audio.

Cost depends on your settings.

Read docs
Input
Advanced settings

Return the final video frame as an image for a follow-up generation.

Reuse a seed with the same inputs to make results more repeatable.

Output
00:00
Example

Try an example

  • Use the reference image as the animal; keep the same snow leopard in every shot. Shot one: the leopard explodes from its crouch on the granite ridge, snow bursting from its paws, camera tracking low and fast alongside through streaming spindrift. Hard cut. Shot two: a wide shot from below as it bounds down the rock face, tail whipping for balance, blue peaks behind. Hard cut. Shot three: crash zoom to its face as it skids to a stop at the cliff edge, breath steaming in the golden rim light. Wind roar, paws crunching snow, a low growl. Photorealistic, no text.

More from ByteDance

Browse allBrowse all models