Reference videos
MP4, MOV, MKV. Up to 3 files, 10MB each, with a combined reference length within 15 seconds.
Reference audio
MP3, WAV, AAC, M4A, OGG. Up to 3 files, 10MB each, with a combined reference length within 15 seconds.
Preview
Preview will appear here
Seedance 2.0 is ByteDance's multimodal AI video model for text-to-video, image-to-video, and reference-driven generation. It combines prompts with images, videos, and audio so you can steer motion, subject consistency, and sound in a more deliberate way.
Generate 4 to 15 second AI videos with one-second control for shorts, ads, product demos, and trailer-style clips.
Choose 480p for faster drafts or 720p when you want cleaner detail for review, pitching, or publishing.
Create landscape, portrait, square, and cinematic formats for YouTube, TikTok, Reels, landing pages, and demos.
Use longer prompts to describe shots, transitions, subject behavior, lighting, sound cues, and reference usage in detail.
Browse output styles that feel closer to real creative work: tighter subject consistency, better camera intent, and richer short-form storytelling.
Cinematic neon action sequence
A futuristic sports car cuts through a neon city at night with fast tracking motion, reflective streets, dramatic lighting, and polished commercial pacing.
Reference-led fantasy worldbuilding
Turn a stylized fantasy environment into a moving shot with drifting leaves, glowing windows, slow camera glide, and layered atmosphere for a richer story feel.
Social ad with stronger subject motion
Create a bright lifestyle clip with energetic camera movement, confident character performance, natural motion, and short-form ad timing designed for social video.
Build from text alone or combine multiple references when the shot needs tighter creative control.
Start from a prompt, then add references only when you need more precise control over subject, motion, or scene direction.
Guide identity, styling, composition, product detail, or environment design with multiple visual references in one run.
Bring in short video references to transfer pacing, movement, transitions, and camera language into the final output.
Upload audio references or turn on native audio generation when sound and motion need to feel more intentionally connected.
Shape output behavior for cleaner iteration, more useful drafts, and smoother downstream editing.
Return the last frame as a still image
Generate synchronized AI audio
Enable optional web search
Best for
Cinematic ads, social clips, concept trailers, product demos, music-led edits, and reference-heavy scenes where motion, consistency, and sound all matter.
Compared with basic text-to-video tools, Seedance 2.0 gives you a wider control surface: multimodal inputs, native audio, longer short-form output, and stronger reference handling. It fits teams that need something closer to a real creative workflow, not just a one-line prompt box.