Seedance 2.0

Seedance 2.0 AI Video Generator

Create multimodal AI video with Seedance 2.0. Combine text, images, video, and audio references to direct subjects, motion, camera language, visual style, and sound—then refine your result with native audio, video editing, and extension.

Text + Image•Video References•Audio References•Native Audio•Video Editing

Released February 2026 • Multimodal Audio-Video Model

Start Creating Explore all Seedance models

Seedance 2.0

Create with Seedance 2.0

Start with text, an image, or supported reference media. Seedance 2.0 is designed for workflows where creative direction comes from more than a prompt alone.

Text

Describe the subject, action, environment, camera, lighting, style, and sound.

Image

Use images to establish characters, products, scenes, composition, or visual direction.

Video

Reference motion, camera behavior, performance, effects, pacing, or an existing sequence.

Audio

Guide voice, rhythm, sound effects, ambience, music, or other audio direction.

Loading your workspace…

Why Choose Seedance 2.0?

Seedance 2.0 is the full general-purpose model of the 2.0 family, built for creators who need stronger quality and control without moving every project into a longer production-oriented workflow.

Multimodal from the start

Combine written instructions with images, video, and audio instead of forcing every creative decision into one prompt.

Direct with references

Use source material to communicate subjects, composition, movement, camera behavior, visual effects, and sound more precisely.

Native audio + video

Create visuals and synchronized audio within one generation workflow.

Edit what already works

Use supported video editing when a strong result only needs a targeted change.

Continue the story

Extend an existing video while directing the next action, camera movement, or story beat.

Choose standard Seedance 2.0 when output quality and broad multimodal control matter more than the speed advantage of Fast or the efficiency advantage of Mini.

See Seedance 2.0 in Action

Seedance 2.0 can move from simple text-driven creation to complex reference-based audiovisual workflows.

Multimodal Video

Direct a Scene with Multiple References

Use different media to define the subject, environment, motion, camera language, visual direction, and sound.

Image to Video

Turn a Reference Image into Motion

Keep the visual foundation while directing action, expression, camera movement, atmosphere, and audio.

Audio + Video

Create Sound with the Scene

Generate dialogue, ambience, effects, music, and visual action as part of the same audiovisual workflow.

What Is Seedance 2.0?

Seedance 2.0 is ByteDance Seed’s multimodal audio-video generation model released in February 2026. It introduced a unified creation architecture that can understand text, images, video, and audio together.

Four input modalities

Use text, images, video, and audio as part of the same creative workflow.

Multiple references

Seedance 2.0 can work with up to 9 images, 3 video clips, and 3 audio clips together with natural-language instructions.

15-second multi-shot video

The model supports high-quality multi-shot audiovisual output up to 15 seconds.

Native audio

Audio and video are generated together, including support for dialogue, effects, ambience, music, and dual-channel audio workflows.

Video editing

Direct changes to an existing sequence instead of regenerating the full creative idea from the beginning.

Video extension

Continue a video with additional action and connected shots based on new instructions.

Complex motion

Seedance 2.0 improves motion stability, multi-subject interaction, physical plausibility, instruction following, and controllability compared with the previous generation.

Create with Text, Images, Video and Audio

A key advantage of Seedance 2.0 is that different types of input can solve different creative problems.

Text

Best for

Story, action, environment, camera, style, and creative intent.

Use text to explain what happens over time. A useful prompt defines the subject, main action, setting, framing, camera movement, visual treatment, and important audio direction.

Images

Best for

Characters, products, locations, composition, costume, and visual style.

Use images when exact appearance matters more than a verbal description. Tell the model what each image represents and which details should remain recognizable.

Video

Best for

Movement, performance, camera language, pacing, transitions, and effects.

A video reference can demonstrate motion or cinematography that would be difficult to describe precisely with words.

Audio

Best for

Voice, music, rhythm, ambience, dialogue, and sound direction.

Use audio when timing or performance depends on sound. Clearly identify what the audio should control in the generated result.

Prompt

Image References

Video References

Audio References

Seedance 2.0

Generated Video

More references are not automatically better. Give each reference a clear role and avoid conflicting creative direction.

From References to Generated Video

Reference media helps turn an abstract prompt into a more specific creative brief. Define what each source contributes, then use the prompt to explain how those elements should work together over time.

Generate Audio and Video Together

Seedance 2.0 treats sound as part of the scene rather than an afterthought.

Dialogue

Generate spoken performance together with the character and scene.

Sound effects

Match environmental or action sounds to what happens on screen.

Ambience

Build environmental sound that supports the visual setting.

Music

Use music or musical direction when it contributes to the scene.

Audiovisual timing

Coordinate visual action with speech, rhythm, effects, or other sound cues.

Native audio can reduce separate production steps, but generated speech, music, effects, and synchronization may still need iteration.

Edit or Continue a Video

Video Editing

Use an existing sequence as a reference and describe what should change. Editing is useful when the overall video works but a subject, action, scene detail, or story element needs revision.

Use editing when

the composition already works
only part of the action needs changing
a subject or scene detail needs revision
regenerating the entire idea would waste a strong result

Video Extension

Continue a sequence beyond its current ending while carrying forward the subject, scene, motion, and narrative direction.

Use extension when

the next shot should follow the existing clip
an action needs another beat
a story should continue
you want connected footage instead of an unrelated generation

Seedance 2.0 vs Fast, Mini and 2.5

The models share related workflows but optimize for different stages of video creation.

Model	Best for	Choose it when	Main trade-off
Seedance 2.0	Full general-purpose multimodal creation	You want strong output quality, native audio, reference control, editing, and extension in the standard 2.0 workflow.	More model than necessary for lightweight drafts or speed-first iteration.
Seedance 2.0 Fast	Rapid iteration	Testing prompts, concepts, revisions, and multiple creative directions quickly.	Optimizes for speed instead of maximizing final-generation quality.
Seedance 2.0 Mini	Efficient higher-volume generation	Creating drafts, social variations, prototypes, and repeated content where efficiency matters.	Less production headroom than the full 2.0 model.
Seedance 2.5	Longer production-focused creation	You need up to 30-second storytelling, stronger references, more precise editing, extension, or advanced production control.	Its additional capability is unnecessary for many standard multimodal tasks.

Seedance 2.0

Best for

Full general-purpose multimodal creation

Choose it when

You want strong output quality, native audio, reference control, editing, and extension in the standard 2.0 workflow.

Main trade-off

More model than necessary for lightweight drafts or speed-first iteration.

Seedance 2.0 Fast

Best for

Rapid iteration

Choose it when

Testing prompts, concepts, revisions, and multiple creative directions quickly.

Main trade-off

Optimizes for speed instead of maximizing final-generation quality.

Seedance 2.0 Mini

Best for

Efficient higher-volume generation

Choose it when

Creating drafts, social variations, prototypes, and repeated content where efficiency matters.

Main trade-off

Less production headroom than the full 2.0 model.

Seedance 2.5

Best for

Longer production-focused creation

Choose it when

You need up to 30-second storytelling, stronger references, more precise editing, extension, or advanced production control.

Main trade-off

Its additional capability is unnecessary for many standard multimodal tasks.

Choose Seedance 2.0 when you need the complete multimodal 2.0 workflow without specifically optimizing for speed, lower-cost volume, or the longer production workflow of 2.5.

Best Use Cases for Seedance 2.0

Cinematic Scenes

Direct camera movement, atmosphere, action, visual style, and synchronized sound in a connected scene.

Character Performance

Combine character references, movement, dialogue, expression, and camera direction in one audiovisual workflow.

Product Advertising

Use product and environment references to explore commercial scenes, demonstrations, reveals, and branded concepts.

Reference-Driven Video

Combine subjects, scenes, camera references, motion examples, and audio to communicate a detailed creative brief.

Previsualization

Test storyboards, blocking, camera language, action, pacing, and sound before committing to a larger production workflow.

Social & Campaign Creative

Develop polished concepts for social content, campaign visuals, short narratives, and creative advertising.

How to Use Seedance 2.0 on DoMax AI

Choose Seedance 2.0

Use the standard model when you need the full multimodal workflow and prioritize quality and control over Fast or Mini optimization.

Choose your starting point

Begin with text, an image, or the reference workflow currently available in the generator.

Add creative references

Upload supported images, videos, or audio when those sources communicate the subject, environment, movement, camera, or sound more clearly than text.

Write the relationship between inputs

Explain which reference defines the character, which defines the scene, which controls movement, and which contributes audio or style.

Generate and refine

Review subject consistency, motion, composition, camera behavior, audio, and instruction following. Refine the prompt or references, or use editing and extension when the base result already works.

Tips for Better Seedance 2.0 Results

Give every reference a job

Do not upload several files without explaining what each one should contribute.

Keep identities consistent

Avoid references that show conflicting character, product, costume, or environment details unless the difference is intentional.

Describe what changes over time

References establish context. The prompt should still explain the action, progression, camera behavior, and intended ending.

Separate appearance from motion

Use visual references for appearance and a motion/video reference for movement when those directions come from different sources.

Control complexity

A clear sequence with connected actions is easier to direct than many unrelated events competing inside the same clip.

Edit strong results

If most of the result works, use editing or extension when available instead of repeatedly rebuilding the entire concept.

Seedance 2.0 Limitations

Complex multi-subject scenes can still show consistency issues, particularly when subjects interact closely or move rapidly.
Fine details, hands, anatomy, small typography, exact product geometry, and highly constrained actions may require additional generations.
Text rendering inside generated video is not guaranteed to remain perfectly accurate.
Complex editing instructions can become less predictable when several subjects, actions, and changes are requested at once.
Native audio is useful, but occasional distortion or imperfect dialogue, music, effects, or synchronization can still occur.
More reference files do not automatically improve the output. Conflicting inputs can make the intended result harder to interpret.
Seedance 2.0 supports up to 15-second high-quality multi-shot audiovisual generation in the official model specification, but exact DoMax duration and generation options must follow the live generator.

More Seedance 2.0 Examples

Explore different ways to combine visual direction, motion, sound, references, and storytelling with Seedance 2.0.

Multimodal

Cinematic

Character

Product

Animation

Advertising

Seedance 2.0 FAQ

What is Seedance 2.0?

Seedance 2.0 is ByteDance Seed’s multimodal audio-video generation model. It supports text, image, video, and audio input and combines reference-based video creation, native audio, editing, extension, and multi-shot generation in one workflow.

Can Seedance 2.0 generate video from text?

Yes. Seedance 2.0 supports text-to-video generation. A prompt can describe the subject, action, environment, camera movement, lighting, visual style, and audio direction.

Can Seedance 2.0 turn an image into video?

Yes. Seedance 2.0 supports image-to-video workflows. Images can define subjects, products, characters, environments, composition, or style while the prompt directs movement and scene development.

What references can Seedance 2.0 use?

Seedance 2.0 supports text, image, video, and audio inputs. ByteDance states that the model can use up to 9 images, 3 video clips, and 3 audio clips together with natural-language instructions.

How long can Seedance 2.0 videos be?

ByteDance’s official Seedance 2.0 specification supports up to 15-second high-quality multi-shot audiovisual output. Exact duration options available on DoMax AI should follow the current generator interface.

Does Seedance 2.0 generate audio?

Yes. Seedance 2.0 uses a unified audio-video generation architecture and can generate sound together with the visuals, including dialogue, ambience, effects, and music-oriented content.

Can Seedance 2.0 edit existing video?

Yes. Seedance 2.0 includes video editing workflows that can use an existing sequence as reference material and apply instructed changes to the result.

Can Seedance 2.0 extend a video?

Yes. Seedance 2.0 supports video extension, allowing a sequence to continue with additional action or story direction while building from the existing video.

What is the difference between Seedance 2.0 and Seedance 2.0 Fast?

Seedance 2.0 is the full general-purpose 2.0 model and is the better choice when output quality has higher priority. Seedance 2.0 Fast is optimized for faster generation and creative iteration.

What is the difference between Seedance 2.0 and Seedance 2.0 Mini?

Seedance 2.0 is intended for higher-quality general-purpose multimodal creation. Seedance 2.0 Mini is the lighter option for efficient drafts, repeated generation, and higher-volume workflows.

Explore More Seedance Models

Longer production

Seedance 2.5

For 30-second storytelling, stronger reference control, advanced editing, extension, and production-focused workflows.

View model

Faster iteration

Seedance 2.0 Fast

For rapid prompt testing, creative variations, and workflows where generation speed is the priority.

View model

Efficient generation

Seedance 2.0 Mini

For drafts, prototypes, repeated creation, and higher-volume workflows.

View model

Audio + video

Seedance 1.5 Pro

For an earlier Seedance workflow focused on native audiovisual generation, dialogue, and character performance.

View model

Explore all Seedance models

Create with Seedance 2.0

Combine text, images, video, and audio to direct complete audiovisual scenes with multimodal references, native sound, editing, and extension.

Start Creating Compare all Seedance models