Reference-Powered Generation
Upload images, video clips, and audio to guide character appearance, camera movement, and scene composition. The model learns from your references, not just your prompts.
Turn references into cinema. Seedance 2.0 unifies characters, motion, and style across shots—giving you the control to direct, not just prompt.
Coming Soon — Multimodal reference engine for consistent characters and cinematic motion
Coming Soon — Multimodal reference engine for consistent characters and cinematic motion
Seedance 2.0 reads images, video clips, and audio references to preserve identity, replicate motion styles, and maintain visual continuity across your entire sequence.
Upload images, video clips, and audio to guide character appearance, camera movement, and scene composition. The model learns from your references, not just your prompts.
Keep the same face, outfit, and brand elements across multiple scenes. Whether your character walks through a city street or a fantasy landscape, their identity stays intact.
Provide a reference clip or image with the look you want—cinematic grading, anime aesthetics, or VFX style—and Seedance 2.0 applies it to your new footage with precision.
Explore what Seedance 2.0 can create across cinematic scenes, stylized portraits, and abstract motion graphics.
Neon-lit megacity at dusk rendered with volumetric lights and flying vehicles.
Prompt
Ultra cinematic futuristic megacity at sunset, neon reflections, volumetric lighting, flying vehicles weaving between skyscrapers, Seedance 2.0
Slow-motion fashion portrait — silk fabric, hair dynamics, and micro expressions.
Prompt
Model's silk gown flowing gracefully, cinematic lighting, subtle micro expressions, generated by Seedance 2.0

Surreal particles, liquid metals, and neon gradients for motion-graphics workflows.
Prompt
Surreal abstract art film with floating shapes, neon gradients, and liquid metal morphing into crystals, Seedance 2.0