Seedance 2.0 Mini is a lightweight video model in ByteDance's Seedance 2.0 series, suitable for short-form previews, character shots, and advertising concept validation. It can generate videos from text or images, and can also combine image, audio, and video references to organize visuals, motion, and rhythm. On this platform, you can select 480p or 720p, landscape or portrait aspect ratios, and audio output to create short shots for clearly defined asset purposes.
Submit requests to the public API at api.acedata.cloud using the documented parameters, then use the results in your application.
Input parameters and result formats vary by service. Use the public API for this model and follow its guide for generation, task retrieval and editing operations.
Clarify capacity, inputs and outputs, and invocation methods before selecting a model.
The above describes this model's creation and invocation scope on this platform; native model specifications and available service parameters should be understood separately.
Learn what doubao-seedance-2-0-mini-260615 can bring to your work.
Multimodal references do more than simply attach an image to a video: reference_image is used for character or subject reference, reference_video for action and camera movement reference, and reference_audio for sound and rhythm reference. Prompts then specify the role each material plays, making this suitable for organizing existing visual assets into new short shots.
When motion needs to begin from a specified image, use first_frame; when the end of the shot needs to be defined, add last_frame. If you want a person or subject to enter a new scene, use reference_image and describe the new environment and action. These two approaches serve different creative goals, so there is no need to mistake subject reference for locking the composition of the original image.
With generate_audio enabled, videos with sound can be generated; prompts can describe both visual actions and the sounds you want to appear. Reference audio, meanwhile, participates in creation as another type of input material. The two have different meanings: providing audio reference is not the same as enabling sound output, so sound goals and reference purposes should be planned separately during creation.
Start with specific tasks to find where the model can be effective.
Input product images and shot descriptions to create different short-film concepts around viewing angles, backgrounds, lighting, and movement. You can first choose 480p to review composition and action, then use 720p to produce selected shots. The deliverable is video material for team review, making it easier to discuss creative directions rather than generating a complete advertising project in one go.
Use clear, authorized images of people or characters as reference_image, together with prompts for scenes, clothing, and actions, to create short shots of the same subject in different environments. Choose 9:16 for vertical content, and clearly specify character behavior and camera movement; after delivery, focus on checking whether the appearance, actions, and scenes match the intended settings.
Use storyboard images as first-frame or first-and-last-frame inputs, describe the motion between images, and generate short shots to demonstrate composition changes. If the focus is on drawing from a particular action or camera movement, use reference video mode instead. Generated results can be used for director communication, animation previs, or concept presentations, helping determine the next production direction.
Choose based on task complexity, input materials, and expected results.
Mini is suitable for lightweight multimodal short-shot creation, especially for tasks where 480p or 720p is sufficient for previews and asset presentation. 2.0 Fast is the fast version in the same series, and both support these two resolutions; if a project explicitly requires higher resolution, consider 2.0 Standard, which supports up to 4k. Do not infer specific processing time or quality differences solely from version names.
For tasks focused on 4–15 second generation, first-and-last-frame control, and image, audio, or video references, choose Mini. If you need up to 30 seconds, creation based solely on audio references, more reference materials, or editing and extending existing videos, Seedance 2.5 is more suitable. These are differences in creation methods, not capabilities that can be obtained by adding fields to Mini.
From a small-scale task to production integration.
Define the goal, required inputs, and output requirements, using real business examples as a starting point.
Open the trial page, confirm the parameters supported by this entry point, then submit a small-scale task to review the results.
Keep the complete model ID, use the request format specified in the documentation, and confirm billing rules on the Pricing page.
Before formal use, understand the output quality and capability scope.
Answers to common questions when using doubao-seedance-2-0-mini-260615.
Submit a request to POST /seedance/videos, set model to doubao-seedance-2-0-mini-260615, and provide the content array. Text items use type=text, with up to 1000 characters per item; it is recommended to specify resolution, aspect ratio, and duration as top-level fields for clear control of generation settings.
Yes. Put the image in the image_url object, specify it as the starting frame with first_frame, and describe the motion and camera movement in text. The image address must be placed in the url field within image_url; you cannot assign a string directly to image_url. For changing the subject's scene, reference_image is more suitable.
Mini supports reference_audio for multimodal reference, but when using audio reference, you must also provide an image or video reference and explain the purpose of each asset in the prompt. Audio reference is not the same as enabling audio output; to generate a video with sound, you must separately set generate_audio=true. If you need to create a video based solely on audio reference, you can choose Seedance 2.5.
Mini's reference_video is used to reference content such as motion and camera movement during generation; it is not the same as a dedicated video editing or extension task. To replace content in an existing video or extend the original clip, use the corresponding mode in Seedance 2.5; Mini is better suited to creating new short videos based on reference materials.
You can set async=true to first obtain a task_id, then query the task result; or provide callback_url to receive a notification when the task is complete. The video link in the completed result is used to download the finished video; do not treat the newly returned task ID as a video file. To get a last-frame image, enable return_last_frame.
Model information · Updated: 2026-10-01. For calling parameters and billing rules, see the API and pricing sections.
Choose a development approach or application suited to this model and follow its setup guide.
↗Python SDK Integration Guide
SDK installation, authentication and examples↗TypeScript SDK Integration Guide
SDK installation, authentication and examples
↗
↗
↗Pricing
$0.0371–$0.20· per second
You pay$13,584for 142,800 credits · $0.0951 / credits46% OFF
You're billed in credits. Pick a top-up tier to see its per-credit rate — the larger the top-up, the cheaper each credit.
This is only a basic call example. See the full docs for more parameters and advanced usage.
View Docs