Turn text, images, video, and audio references into controlled AI videos with cinematic motion, camera direction, consistent visuals, and synchronized sound.
Seedance 2.0 is a multimodal AI video generation model from ByteDance designed to work with text, images, videos, and audio as creative inputs. Instead of relying on a text prompt alone, you can use reference assets to guide characters, scenes, motion, camera movement, visual style, and sound.
The model can generate short videos, work from image and video references, synchronize visuals with audio, and support workflows such as video extension and targeted editing. This makes it useful for creators, marketing teams, e-commerce brands, and other users who want more direct control over AI-generated video.
Use references instead of trying to explain every visual detail in a long prompt.
Build your video from multiple types of input at once. Seedance 2.0 supports up to nine image references, three video clips, and three audio files within a mixed-input workflow. Use one asset to establish the character, another for the scene, a video for camera movement, and audio for timing or atmosphere.
Tell the model what each uploaded asset should influence. Use references for character appearance, clothing, environments, visual effects, movement, camera direction, sound, or overall visual treatment.
Use reference video to communicate movement visually. This can help with choreography, action timing, tracking shots, camera rotation, transitions, and other motion that is difficult to describe precisely with text alone.
Reference images and videos can help preserve characters, clothing, scene design, products, and overall visual style. Results can still vary, especially in complex scenes, so important commercial outputs should be reviewed before use.
Seedance 2.0 supports audio-visual generation, including dialogue, sound effects, music, and lip-sync workflows. Uploaded audio can also guide timing, beats, pacing, or atmosphere.
Continue an existing sequence instead of starting over. Supported workflows can include video extension, connecting clips, replacing selected elements, adding content, or modifying actions while preserving other parts of the scene.
Explore example Seedance 2.0 generations, then reuse a prompt with your own references to create a new version.
You do not need to describe everything from scratch. Start with the assets that already communicate the look, motion, or sound you want.
Add images, videos, or audio files that represent your characters, product, scene, movement, visual style, camera behavior, or soundtrack. You can combine multiple input types in the same generation.
Write a natural-language prompt explaining the subject and action, scene, camera movement, pacing, lighting, visual style, audio, and how each reference should be used.
Select the duration, output resolution, aspect ratio, and other available settings. Use 9:16 for vertical social content or 16:9 for widescreen scenes.
Generate your first version, review the result, and adjust only what needs improvement. Refine the prompt, change references, extend the sequence, or make targeted edits.
“Use @image1 as the main product reference. Start with a close-up, then follow the slow orbiting camera movement from @video1. Match the lighting to @image2 and time the final reveal to the beat in @audio1.”
Combine text, image, video, and audio references in a single workflow.
Blend multiple reference types into one generation.
Guide characters, scenes, products, and visual style.
Guide motion, choreography, and camera movement.
Guide timing, beats, pacing, and atmosphere.
Short-form clips per generation.
Standard output options for Seedance 2.0.
Cinematic 24 fps output.
Widescreen, vertical, square, and other formats.
Dialogue, sound effects, music, and lip-sync support.
Describe the video in natural language.
Extend sequences and make targeted modifications.
Use these structures as starting points and replace the references with your own assets.
“Use @image1 as the exact product reference. Place it on a reflective studio surface with soft side lighting. Start with a macro detail shot, pull back slowly, then rotate around the product. Add subtle cinematic sound effects.”
Use prompt“Use @image1 for the character's appearance and clothing. Keep the same character throughout the sequence. Follow the walking motion from @video1 and place the character in the environment shown in @image2.”
Use prompt“Recreate the camera movement from @video1 in a new nighttime city scene. Begin behind the subject, move into a side tracking shot, and finish with a slow push-in.”
Use prompt“Create a vertical 9:16 product video using @image1 as the product reference. Use quick cuts, close-up details, and the rhythm of @audio1. End on a clean hero shot.”
Use prompt“Use @audio1 as the timing reference. Create a cinematic performance scene with cuts and camera movement aligned to major beats. Maintain the same character and lighting style throughout.”
Use prompt“Continue the uploaded clip naturally. Preserve the subject, environment, camera direction, lighting, and movement. Extend the action into a wider final shot.”
Use promptTurn ideas, reference images, music, and existing clips into TikTok videos, Instagram Reels, YouTube Shorts, music visuals, personal vlogs, dance content, cinematic experiments, and other short-form formats.
Use products, campaign assets, creative references, and motion examples to develop product videos, campaign hooks, brand content, social ads, creative variations, and pre-visualization faster.
Turn product imagery and brand assets into product demos, launch concepts, Shopify visuals, virtual spokesperson scenes, product reveals, lifestyle content, try-on concepts, and branded social assets.
Use reference-driven generation for storyboarding, pre-visualization, camera planning, motion tests, VFX concepts, architectural visualization, training materials, and early-stage creative exploration.
Check the latest Seedance 2.0 generation cost, free credits, and plan requirements before you create.
600 credits · $0.062 per credit
Perfect for trying out PWZ — get started with a credit balance for your first generations.
2,000 credits · $0.052 per credit
For creators generating regularly — better per-credit value for daily PWZ workflows.
4,000 credits · $0.045 per credit
Best value for teams and content pipelines — more generations at the lowest per-credit cost.
Seedance 2.0 is a ByteDance AI video generation model that supports text, image, video, and audio inputs. It is designed for reference-driven video generation, motion and camera guidance, audio-visual creation, and video editing workflows.
Seedance 2.0 supports text prompts plus image, video, and audio references. The supplied specifications list up to nine images, three video clips, and three audio files, with a maximum of 12 mixed files.
The supplied Seedance 2.0 specifications list generation lengths from 4 to 15 seconds per shot. Longer sequences may require extension or multiple connected generations.
The supplied specifications list 480p, 720p, and 1080p output. Any 2K or 4K option shown by a provider should be checked to determine whether it is native generation, upscaling, or platform-specific processing.
Yes. You can upload images as references for characters, products, scenes, composition, or visual style and use a text prompt to describe the motion and result you want.
Yes. Reference video can be used to guide motion, choreography, action, transitions, and camera movement. The supplied specifications list support for up to three video clips.
The supplied materials describe audio generation including dialogue, sound effects, music, and lip-sync. Audio files can also be used as references for timing, beats, and atmosphere.
Yes, the supplied capability information includes video extension and targeted editing workflows. Available editing controls may depend on the way Seedance 2.0 is integrated into PWZ.
Bring your prompt, images, clips, and audio together with Seedance 2.0. Guide the scene with real references, generate your first version, and refine it from there.