What Is Seedance? Models, Inputs, and Video Costs

What Seedance is

Seedance is ByteDance Seed's family of video generation models. It can turn written directions and supported image, video, or audio references into a video with sound. ByteDance introduced Seedance 2.0 in February 2026 and announced Seedance 2.5 in July 2026. This site, SeedanceAI at seedanceai.im, is an independent third-party tool; it is not ByteDance's official product.

The model makes the clip. The workspace on this site supplies a prompt editor, reference uploads, a credit quote before submission, and a place to review the resulting job. A model provider may reject an unsupported combination even when the workspace accepts the files, so check the result status and error message.

Choose a starting point

Text to video: Describe the subject, what changes during the shot, the camera movement, and any sound. For example: “At dusk, follow a cyclist along a wet street in a steady side tracking shot. Keep the rider in frame as the shop lights pass behind them; hear light rain and tires on pavement.” Select Text to Video when you have no starting image.

Image to video: Upload an opening image and describe its movement. An optional last frame can guide where the scene ends. The prompt should explain the action between the frames rather than simply repeating what is already visible.

Multi Reference: Combine supported media, then assign a role to each file in the prompt. A photo might establish the subject, a video might guide a camera move, and audio might indicate the sound. The workspace's @ picker helps refer to each uploaded asset. Only upload material you have permission to use.

Seedance 2.0 and 2.5 on this site

Workspace setting	Seedance 2.0	Seedance 2.5
Output length	4–15 seconds	4–30 seconds
Output resolution	480p, 720p, 1080p	480p, 720p, 1080p
Maximum reference files	9 images, 3 videos, 3 audio clips	30 images, 10 videos, 10 audio clips
Total video and audio reference length	15 seconds each	30 seconds each
Draft and video editing in this workspace	No	480p draft; edit or extend from a source video

These are the limits currently checked by this workspace. ByteDance's Seedance 2.0 announcement and Seedance 2.5 announcement describe the model families. Provider availability and individual input combinations can differ. For the exact form and workflow, open the Seedance 2.0 generator or Seedance 2.5 generator.

What a generation costs

Credits depend on the model, output length, resolution, batch size, and reference media. For one 5-second, 720p output without reference video or audio, this site's current quote is 15 credits on Seedance 2.0 and 20 credits on Seedance 2.5. The 480p and 1080p prices differ. Reference video length changes the base calculation; adding reference video or audio doubles the calculated amount once. A draft and its final version are separate requests with separate charges.

The generator shows the actual quote for your selected settings before submission. See the Seedance pricing table for each model and resolution. There is no guaranteed signup credit allowance; check your available balance before generating. You can read the sample prompts and view the illustrative gallery without spending generation credits. The gallery's in-house motion studies are not evidence of output from either model.

Which model should you try?

Start with Seedance 2.0 when the scene fits within 15 seconds and you want to compare a lower-cost 720p base quote. Choose Seedance 2.5 when the shot needs more time, a 480p draft workflow, or the workspace's edit and extend options. Keep the same prompt, duration, resolution, and sound setting when comparing models; otherwise the two results do not isolate the model choice. See the 2.5 settings and cost guide for a reproducible comparison procedure.