Seedance

Seedance 2.0 AI Video Generator

What Is Seedance 2.0?

Seedance 2.0 is ByteDance’s video generation model for short scenes built from text, images, video, and audio. Use it for character-led stories, product films, music videos, or shot previews before a larger production.

Connected Stories in 4–15 Seconds

A car film can move from a wide view of the landscape to a tracking shot and a final detail. Each shot has a purpose within one short sequence.

Realistic Characters and Consistent Details

In a fashion film, changes in camera angle reveal the outfit, expression, and movement. The subject remains the focus as the scene shifts from an establishing view to a close-up.

Physical Motion and Coordinated Action

An animated fight needs a clear sequence of attack, contact, and response. Use a short action scene to explore timing and weight before developing a longer sequence.

Native Sound, Lip Sync, and Musical Timing

A band performance brings voices, instruments, and movement into the same scene. It gives you a starting point for a music-video shot or a performance-led advert.

Seedance 2.0 Features

Direct the performance, movement, editing, and sound of a short scene with text and reference material.

Official demo · From reference images and a shot plan to a connected scene.

Four Types of Input

Combine text with up to nine images, three videos, and three audio clips. Images guide appearance, videos guide motion or effects, and audio guides dialogue and rhythm. Specify the role of each reference in the prompt.

Camera Movement and Scene Direction

Reference footage can guide tracking, orbits, push-ins, composition, blocking, and movement. Apply that camera plan to your own subject and setting.

Visual Effects and Transitions

Use footage to guide effects, whip pans, match cuts, reveals, and editing rhythm in a new scene.

Storyboards into Connected Shots

Turn a storyboard or written shot plan into a connected sequence, linking planned scenes and actions so changes in framing advance the story.

Character Performance and Continuity

References help preserve a character’s appearance, clothing, or product details across camera angles. Facial expressions, gestures, fabric, and lighting contribute to the performance.

Weight and Motion in Complex Scenes

Generate actions involving gravity, momentum, contact, and timing, including coordinated movement, objects responding to forces, and fabric moving with a subject.

Native Stereo Audio and Lip Sync

Generate dialogue, ambience, music, and effects with the picture, including stereo sound. Lip movement follows speech, while footsteps and impacts accompany visible actions.

Music-Guided Motion and Cuts

An audio reference can guide the timing and mood of actions, cuts, and transitions. Build a dance, reveal, or scene change around a musical phrase.

Targeted Changes and Video Continuation

Adjust a person, action, or story detail using reference footage, or continue an existing clip with a subsequent action. Useful for revising a moment or adding a shot.

Who Is Seedance 2.0 For?

Content Creators and Marketing Teams
Test product adverts and social clips before committing to a production direction.
Filmmakers and Video Editors
Preview performances, camera positions, and scene transitions from a shot plan.
Animation and Game Teams
Explore character movement, composition, and action timing before production.

What Can You Make with Seedance 2.0?

Seedance 2.0 vs. Sora 2 vs. Veo 3

CriterionSeedance 2.0Sora 2Veo 3 (3.0)
Reference InputsText, images, video, and audio together.Text and first-frame images; separate video-based character setup.Text and images; no video or audio references in 3.0.
Shot ContinuityCharacter or product references across connected shots.Multi-shot instructions and persistent scene state.Opening image plus subject and scene prompts.
Camera and MotionVideo references guide movement, framing, and transitions.Actions and camera movement described in prompts.Actions and camera movement described in prompts.
SoundNative dialogue, music, and effects; audio references.Synchronized dialogue and sound effects.Sound generated from the audio instructions in a prompt.
Editing and ContinuationTargeted changes and continuation from an existing clip.Documented support for controlled edits and extensions.Video extension added in 3.1; unavailable in 3.0.
AvailabilityAvailable here; 4–15 seconds per generation.Official apps and API discontinued in 2026.Google Cloud 3.0 endpoints retired; 3.1 recommended.

Seedance 2.0: Strengths and Limits

Reasons to Choose It

  • Separate roles for each reference: Guide appearance, movement, and sound with the material best suited to each.
  • Connected shots: Use a shot plan and character references to build a short sequence.
  • Picture and sound together: Review dialogue, music, and effects alongside the generated video.

Limits to Consider Before Creating

  • 4–15 seconds per generation: Longer videos need separate scenes and editing.
  • Clear direction still matters: Give each reference a purpose and review complex actions shot by shot.

How to Use Seedance 2.0

Start with text or reference material, then generate and download your video on this page.

  1. Choose How to Start

    Seedance 2.0 is already selected. Use Text to Video for a written idea, or switch to Image to Video and upload your references.

  2. Describe the Scene

    Write the subject, action, and camera movement. Set duration, aspect ratio, and resolution; enable audio for dialogue or sound.

  3. Generate and Download

    Check the credit cost, sign in, and submit. Preview and download the finished clip, or return to saved results in My Creations.

Explore more AI models

Sunlit stone courtyard with olive trees and reflecting water
Video model

Seedance 2.5

Explore longer scenes and precise reference-based video creation.

Explore model
Red kite above a miniature moss-covered island
Video model

Seedance 2.0 Mini

Plan short concepts and compare creative directions.

Explore model
Yellow sport watch with woven strap on gray stone
Image model

Nano Banana 2

Create and edit the reference images for your next scene.

Explore model
Red ceramics poster with CLAY typography
Image model

GPT Image 2.5

Prepare layouts and illustrations for a visual starting point.

Explore model

Frequently Asked Questions

Answers about costs, output settings, reference limits, model versions, and saved videos.

Yes. Use Text to Video for a written scene, or Image to Video for first and last frames or mixed image, video, and audio references. First/last frames and mixed references are separate input methods in this workspace.

Use up to 9 images, 3 videos, and 3 audio clips. Video references have a combined 15-second limit; audio references have a separate 15-second total. Each video file must be no larger than 50 MB. State whether each reference should guide appearance, action, camera movement, or sound.

One generation lasts 4–15 seconds. The standard model’s resolution menu here lists 480p, 720p, 1080p, and 4K; Fast lists 480p and 720p. Aspect ratios include 16:9, 9:16, 4:3, 3:4, 1:1, 21:9, and Adaptive. Check the selected settings and credit cost before submitting.

In Image to Video, first and last frames guide the opening and closing images; the prompt describes the movement between them. Use this method separately from mixed references. Return Last Frame is an output option that retrieves the generated clip’s final frame for use as a starting point for another shot.

The standard model focuses on detailed scenes and output quality. Fast is the accelerated variant for testing ideas and making revisions. Both accept text, image, video, and audio references, but their resolution menus differ here. Compare the available settings and credit cost in the model selector.

Seedance 2.0 generates 4–15 second clips and accepts up to 9 images, 3 videos, and 3 audio clips. Seedance 2.5 supports clips up to 30 seconds and larger reference sets. Choose according to the length of the scene and the amount of reference material you need.

The generator displays the credit cost for your selected settings before submission. Check it again after changing the model version, duration, resolution, or other settings; the cost can change with those choices.

Preview and download the finished video in the workspace. Signed-in users can also return to saved results in My Creations to watch or download them again.
Rust-red desert ridges beneath a deep blue twilight sky

Create with Seedance 2.0 on Seedance

Create with prompts and references in one workspace. See the credit cost before generating, then preview, download, and return to your saved videos.

Open My Creations
Create a Reference Image