Text to Video
Write a subject, one action, camera direction and sound in a compact prompt. Text mode takes no uploaded media, making it the simplest way to compare two prompts with the same settings.
Seedance 2.0 AI Video Generator
Create up to 15 seconds of video from text, a first frame, or multiple references. Describe the action and camera movement, then guide the result with images, video, or audio.
What is Seedance 2.0?
Seedance 2.0 creates short clips from a written direction, a starting image, or a combination of media references. Use it to test the movement, framing, and sound of a scene before developing the next shot.
Choose a starting point for a short scene
Seedance 2.0 makes one continuous clip from a written scene, a starting image, or supported reference media. Choose the mode before writing the prompt so each input has a clear purpose.
Image to Video
Upload a first frame, then describe what should move. An optional last frame can guide the ending. This mode accepts images only, so switch modes if you need video or audio references.
Multi Reference
Combine an image or video with other supported references. Use the @ picker in the prompt to assign a role to each asset; the selected model limits file counts and total media duration.
Seedance 2.0 settings in this workspace
These limits are checked before a request is sent. The provider can still reject a particular combination of otherwise valid inputs.
| Setting | Available option |
|---|---|
| Output | 4–15 seconds; 480p / 720p / 1080p |
| Reference files | Up to 9 images, 3 videos and 3 audio clips |
| Reference length | Up to 15 seconds of video and 15 seconds of audio in total, counted separately |
| Operations | Generation only; draft preview, video edit and video extend require Seedance 2.5 |
Image to Video uses one first-frame image and at most one last-frame image. Multi Reference needs at least one image or video. The workspace shows current options for the selected mode.
What five and fifteen seconds cost on Seedance 2.0
Credits for one output. The first two columns use text mode with no reference media. The final column uses one five-second reference video with a five-second output; no reference audio is included.
| Resolution | 5s text prompt | 15s text prompt | 5s + 5s video reference |
|---|---|---|---|
| 480p | 10 | 30 | 20 |
| 720p | 15 | 45 | 40 |
| 1080p | 30 | 90 | 80 |
A video reference adds its length to the calculation, then doubles the calculated credits. Audio references also apply the multiplier, but using both kinds still doubles only once. The generator quotes the exact settings before you submit.
Run your own same-prompt comparison
This is a reproducible test plan, not a claim about outputs. The illustrative clips elsewhere on this site are not model test results.
- 1
Hold the scene constant
In text mode, set one output at five seconds and 720p. Try: ‘A red kite rises above a calm beach at dawn. In one continuous wide shot, the camera follows it upward while waves roll below.’
- 2
Record both quotes
With no references, Seedance 2.0 quotes 15 credits for that setting. Switch to 2.5 without changing the prompt, duration or resolution; it quotes 20 credits. Record each setting before submitting.
- 3
Compare only actual results
If you run both generations, save their prompt, model, settings and result status. Compare subject consistency, camera path and timing from the resulting clips rather than these illustrative samples.
When a Seedance 2.0 request does not complete
Use the workspace error and status to decide what to change; these are input checks and recovery guidance, not a log of measured model failures.
Check the mode and limits
Keep output between four and fifteen seconds. Draft, Edit and Extend belong to 2.5. Text mode takes no files, Image mode takes one or two images, and Multi Reference needs an image or video within the file and duration limits.
Check the existing result
A processing request may still finish. Read its status before starting another generation. If the provider confirms failure, charged credits are returned; insufficient balance prevents submission in the first place.
Explore Another Seedance Model
Seedance 2.5Seedance 2.0 MiniCreate Your Next Scene with Seedance 2.0
Start from a prompt or add your reference media. Choose a Seedance model and direct the shot.
Start Creating Videos