Prompt to cinematic clip
Describe the scene, camera path, subject action and lighting to generate a polished video from text alone.
Create cinematic videos from prompts, keyframes, reference images, video clips and audio cues with Seedance 2.0.
Generate a video directly from a detailed prompt.
Preview the Seedance 2.0 style, then generate your own clip from text prompts, image keyframes or multi-reference media.
Seedance 2.0 turns prompts, first and last frames, real-person reference images, short video references and audio cues into polished AI video clips with stronger subject consistency. Use Fast for quick iteration or Quality for richer 1080p output.
Seedance 2.0 is built for flexible video direction: prompt-only generation, frame-controlled motion, real-person reference guidance, broad aspect ratios and optional audio.
Start from a prompt, anchor the first and last frame, or guide people, style and motion with authorized reference images, videos and audio.
Use wide, vertical, square, cinematic 21:9 or adaptive aspect ratios while describing camera movement, lens feel and scene rhythm.
Turn on audio generation when the scene needs environmental sound, action accents or richer social-ready output.
Choose Seedance 2.0 Fast for quick drafts, or Seedance 2.0 Quality for 1080p-capable final clips.
Use Seedance 2.0 for product motion, cinematic ads, storyboards, social clips, music-led edits and real-person reference scene exploration.
Describe the scene, camera path, subject action and lighting to generate a polished video from text alone.
Upload keyframes to define the start and finish, then let Seedance fill the motion between them.
Use authorized human reference photos or clips to help preserve a person’s look, styling and role while changing camera movement, scene or mood.
Detailed answers about Seedance 2.0 modes, model selection, reference inputs, pricing, history and generation quality.
Use the generator above to create text-driven, keyframe-guided or reference-rich videos with Seedance 2.0.
Back to generator