Multimodal Reference

Model

Reference images0/9

Reference videos(max 3, total 15s)0.0s/15s

Up to 3 clips (MP4, WebM, MOV). Max 50MB each. Total duration ≤ 15s.

Reference audios(max 3, total 15s)0.0s/15s

Up to 3 tracks (MP3, WAV, AAC). Max 20MB each. Total duration ≤ 15s. Pair with image or video — not text alone. Audio is free and does not consume credits.

Prompt

0/5000

Aspect ratio

Duration

4s10s15s

Resolution

Estimated credits

100

cinematic

"Cinematic AI-generated scene with stunning visual quality and natural motion."

AI-Powered

Create with Multimodal Reference

Guide Seedance 2.0 with images, video clips, audio, and text prompts to lock in motion, style, and sound in one workflow.

20 free credits

Examples

Style Transfer Animation

Apply neon cyberpunk style to this walking person reference video

Character Replacement

Replace the person in the reference video with a robot character, keeping the same motions

Environment Modification

Change the background of the reference video from day to night with rain

Features

Upload up to 9 reference images, 3 videos, and 3 audio tracks
Use natural language to reference specific materials with @ prompts
Available only on Seedance 2.0 and Seedance 2.0 Fast
Maintain motion, subject, and color consistency across generations
Reference-video jobs use the `+video` rate over the total billable duration; audio is free
Generate cinematic variants without switching tools

Powered by Leading AI Models

Seedance 2.0

ByteDance

Seedance 2.0 Fast

ByteDance

Ready to create amazing videos? Sign up now and get started!