Seedance Video Generation

The ID of the model you need to call (Model ID)

The callback notification URL for the generation task results. When the video generation task status changes, the platform will send a POST request to this URL.

The callback request body structure is the same as the query task API response.

The callback status includes the following states:

• queued: In queue.
• running: Task running.
• succeeded: Task successful. (If sending fails, i.e., no successful delivery within 5 seconds, retry up to 3 times)
• failed: Task failed. (If sending fails, i.e., no successful delivery within 5 seconds, retry up to 3 times)
• expired: Task timed out, meaning the task has been in running or queued state beyond the expiration time. Can be configured via the execution_expires_after field.

true: Return the last frame image of the generated video. When set to true, you can obtain the last frame image through the query video generation task API. The last frame image is in PNG format with the same pixel dimensions as the generated video, without watermarks.

Use this parameter to generate multiple consecutive videos: use the last frame of one video as the first frame of the next video task to quickly generate a sequence of consecutive videos.

false: Do not return the last frame image of the generated video.

The service tier of a submitted task cannot be modified Seedance 2.0 series only supports online inference mode and does not support this parameter Currently offline inference is not yet supported

Specifies the service tier for processing this request. Enum values:

• default: Online inference mode, lower RPM and concurrency quotas, suitable for scenarios requiring low-latency inference.
• flex: Offline inference mode, higher TPD quotas, priced at 50% of online inference, suitable for scenarios with less stringent latency requirements.

Task timeout threshold. Specifies the expiration time after task submission (in seconds), calculated from the created_at timestamp. Default: 172800 seconds (48 hours). Range: [3600, 259200].

Regardless of the service_tier used, it is recommended to set an appropriate timeout based on your business scenario. After this time, the task will be automatically terminated and marked as expired.

Only supported by Seedance 2.0 series and Seedance 1.5 pro

Controls whether the generated video contains synchronized audio.

• true: The model outputs video with synchronized audio. The model automatically generates matching voice, sound effects, and background music based on the text prompt and visual content. It is recommended to place dialogue in double quotes for better audio generation. For example: The man stops the woman and says, "Remember, never point your finger at the moon."
• false: The model outputs video without audio.

Note: Generated audio videos are all mono, regardless of the input audio channel count.

Only supported by Seedance 1.5 pro

Controls whether to enable draft mode.

• true: Enable draft mode to generate a preview video, quickly verifying whether the scene structure, camera movement, subject actions, and prompt intent meet expectations. Consumes fewer tokens than normal video, reducing cost.
• false: Disable draft mode and generate a normal video.

Note: When draft mode is enabled, Draft videos are generated at 480p resolution (using other resolutions will cause an error). The last frame return feature is not supported, and offline inference is not supported.

Unique identifier of end users, used to help the platform detect users in your application who may violate the ModelArk usage policy. This identifier is an English string, which must be fixed and unique for a single user, and the length cannot exceed 64 characters. It is recommended to pass in a string generated by hashing the username, user ID or email address to avoid leaking user privacy information.

Only supported by Seedance 2.0 series. (Please contact sales for the maximum priority available per user)

Sets the execution priority of the current request, determining its position in the queue. Range: 0~9, higher values indicate higher priority.

By default, requests are executed in FIFO (First In, First Out) order. When a higher priority is set, the request will jump ahead of all lower-priority requests under the same Endpoint.

Notes:

• Requests with the same priority are still ordered by FIFO.
• Priority only affects queue order and does not interrupt currently running (status=running) tasks.
• Priority only takes effect within the same Endpoint and does not affect other Endpoints.
• Offline inference mode (service_tier=flex) does not support priority configuration.

Example:

An Endpoint currently has 3 queued (status=queued) tasks in the queue, all with priority 0 (default).

Queue: [Task A: priority=0] → [Task B: priority=0] → [Task C: priority=0]

When a new request with priority=5 is submitted, it goes directly to the front of the queue:

Queue: [New Request: priority=5] → [Task A: priority=0] → [Task B: priority=0] → [Task C: priority=0]

Seedance 2.0 series and Seedance 1.5 pro default: 720p Seedance 1.0 pro & pro-fast default: 1080p

Video resolution. Enum values:

• 480p
• 720p
• 1080p: Not supported by Seedance 2.0 Fast and Seedance 2.0 Mini.
• 4k: Only supported by Seedance 2.0

- Compared to typical 8-bit color depth, Seedance 2.0 4k output uses 10-bit encoding, fully preserving rich color gradations and smooth transitions, meeting professional film production and HDR video content requirements.
- 4k videos are output in H.265 (HEVC) format; some players/browsers may not support direct playback.

Seedance 2.0 series and Seedance 1.5 pro default: adaptive Other models: Text-to-Video default 16:9, Image-to-Video default adaptive

The aspect ratio of the generated video. See the table below for pixel dimensions for different aspect ratios.

• 16:9
• 4:3
• 1:1
• 3:4
• 9:16
• 21:9
• adaptive: Automatically selects the most suitable aspect ratio based on input (see details below)

adaptive Rules

When ratio is set to adaptive, the model automatically adjusts the aspect ratio based on the generation scenario. The actual video aspect ratio can be obtained from the ratio field returned by the query video generation task API.

****Supported models:

• Seedance 2.0 series, Seedance 1.5 Pro supported
• Other models only supported in Image-to-Video scenarios

****Rules:

• Text-to-Video：Based on the input prompt, intelligently selects the most suitable aspect ratio.
• First Frame / First & Last Frame: Automatically selects the closest aspect ratio based on the uploaded first frame image ratio.
• Multimodal Reference-to-Video: Determines based on user prompt intent. If it is first-frame video/editing video/extending video, selects the closest aspect ratio based on that image/video; otherwise, selects the closest aspect ratio based on the first media file (priority: video > image).

Pixel dimensions corresponding to different aspect ratios

Note: For Image-to-Video, if the selected aspect ratio does not match the uploaded image ratio, the platform will crop your image from the center.

Video duration, integer only, in seconds.

• Seedance 1.0 pro, Seedance 1.0 pro fast: [2, 12] s.
• Seedance 1.5 pro: [4, 12] or set to -1
• Seedance 2.0 series: [4, 15] or set to -1

Note

Seedance 2.0 series and Seedance 1.5 pro support two configuration methods:

• Specify exact duration: Any integer within the valid range.
• Smart mode: Set to -1 to let the model choose the appropriate video length (in whole seconds) within the valid range. The actual generated video duration can be obtained from the duration field returned by the query video generation task API. Note that video duration is related to billing, so set it carefully.

Temporarily not supported by Seedance 2.0 series and Seedance 1.5 pro Choose either duration or frames; frames takes priority over duration. If you want to generate videos with fractional seconds, it is recommended to specify frames.

The number of frames for the generated video. By specifying the frame count, you can flexibly control the video length and generate videos with fractional seconds.

Due to frames value constraints, only limited fractional seconds are supported. You need to calculate the closest frame count using the formula.

• Formula: frames = duration x frame rate (24).
• Range: Supports all integer values in the [29, 289] range that satisfy the format 25 + 4n, where n is a positive integer.

Example: To generate a 2.4-second video, frames = 2.4 x 24 = 57.6. Since frames does not support 57.6, you can only choose the closest value. According to 25+4n, the closest frame count is 57, and the actual generated video is 57/24 = 2.375 seconds.

Temporarily not supported by Seedance 2.0 series

Seed integer for controlling the randomness of generated content. Range: integer in [-1, 2^32-1].

Notes:

• With the same request, if the model receives different seed values (e.g., not specifying a seed, setting seed to -1 which uses a random number, or manually changing the seed value), different results will be generated.
• With the same request, if the model receives the same seed value, similar results will be generated, but exact consistency is not guaranteed.

Not supported in reference image scenarios, temporarily not supported by Seedance 2.0 series

Whether to fix the camera.

• true: Fix the camera. The platform will append fixed camera instructions to the user prompt, but the actual effect is not guaranteed.
• false: Do not fix the camera.

Whether the generated video contains a watermark.

• false: The generated video does not contain a watermark.
• true: The generated video displays an "AI Generated" watermark in the bottom-right corner.

Video generation task ID. Only stored for 7 days (from the created_at timestamp), then automatically deleted.

• When draft: true, this is the Draft video task ID.
• When draft: false, this is the normal video task ID.

Video generation is asynchronous. After obtaining the ID, use the query video generation task API to check the status. When successful, the API returns the generated video URL.