Text to video
Describe a scene and generate a clip, with people and motion that tend to look natural.
HomeSeedance 2.0
Seedance 2.0
Seedance 2.0 - Multimodal video generation model with text, image, video, audio, first-last-frame, synced audio, and realistic human video support
Main ways to use Seedance 2.0
Reference-driven video
Bring in an image, a video, or audio as a reference to guide the look, motion, or timing of the result.
What Seedance 2.0 does well
Realistic human motion
It handles people — movement, gesture, and presence — more convincingly than many video models, which suits talking or performing subjects.
Multimodal references
Combine image, video, and audio references to steer how a scene looks and moves.
Frame and reference control
Use a starting image or first/last frames to keep the result close to what you have in mind.
How to use Seedance 2.0 on DreamWith
Open the video generator
Sign in to DreamWith, open the video generator, and pick Seedance 2.0 from the model list.
Write your prompt and add references
Describe what you want in plain language. Upload a reference when you need image-to-video, a starting frame, or a look to follow.
Set the parameters
Choose the aspect ratio, duration and resolution, and any other options Seedance 2.0 supports. DreamWith shows the credit cost before you run the task.
Generate, compare, and reuse
Run the generation, compare versions, then refine the prompt or references. Every result stays in your history for reuse.
Where Seedance 2.0 fits best
People-centred content
Create clips where a person speaks or performs and needs to look natural.
Reference-matched scenes
Match a look or motion you already have by feeding it as a reference.
Social and ad video
Produce short clips for feeds and campaigns that feature people.
Why run Seedance 2.0 on DreamWith
One workspace, many models
Use Seedance 2.0 next to other image and video models without juggling separate sites, logins, or balances.
Test before you commit
Compare prompts, references, and settings cheaply before you spend time and credits on final assets.
Shared credits and history
Credits work across every model, and your generation history keeps prompts and outputs together.
FAQs about Seedance 2.0
What is Seedance 2.0?+
Seedance 2.0 is a multimodal video model on DreamWith that accepts image, video, and audio references and is strong at realistic human video.
What references can it use?+
It can take image, video, and audio references, along with text prompts and first/last-frame control.
Is it good for videos with people?+
Yes. Natural human motion is one of its strengths, which makes it a good pick for talking or performing subjects.
How much does it cost?+
The credit cost depends on duration and settings and is shown in the generator before you run a task.
Model keywords
Seedance 2.0 model workflows and related AI models
DreamWith supports Seedance 2.0 for the launched model workflows listed below, with direct paths to compare adjacent models in the same workspace.
Seedance 2.0Seedance 2.0 AI video generatorSeedance AI video generatorSeedance text-to-videoSeedance image-to-videoSeedance 2.0 text-to-videoSeedance 2.0 image-to-videoSeedance 2.0 reference-to-video
AI Video GeneratorOpen the video generation workspace and compare launched video models.Veo 3.1 FastExplore Veo 3.1 Fast keywords, settings, and generation workflows.Veo 3.1 QualityExplore Veo 3.1 Quality keywords, settings, and generation workflows.Seedance 2.0 FastExplore Seedance 2.0 Fast keywords, settings, and generation workflows.Kling 2.6Explore Kling 2.6 keywords, settings, and generation workflows.