Skip to main content
POST
Wan 3.0 - Video
Generate videos with the Wan 3.0 standard video model. Supported generation modes:
  • Text to video
  • First-frame to video
  • First-and-last-frame to video
  • Reference to video
  • File to video
  • Link to video

Create a Video Generation Task

Authentication

All API requests require a Bearer Token.

Model

Use:

Text to Video

First Frame to Video

First and Last Frames to Video

Reference to Video

Reference media can include images, videos, and audio. In the prompt, reference items correspond to their array order:
  • Image1, Image2, …
  • Video1, Video2, …
  • Audio1, Audio2, …

File to Video

The referenced webpage must be publicly accessible and must not require login.

Parameters

string
required
Must be wan/3-0-video.

input Parameters

object
required
The input object contains the parameters used to generate the video.
string
Use prompt to describe the video motion you want to generate in English or Chinese.Note: Required for text-to-video.Maximum length: 20,000 characters.
string
Use the provided image URL as the first frame.
The first_frame_url must contain the URL of an uploaded image. The value must be an uploaded file URL, not the file content itself.
Note: first_frame_url cannot be used together with reference_*_urls according to the source specification.Supported image types:
  • image/jpeg
  • image/png
  • image/webp
  • image/bmp
Maximum file size: 20 MB
Each side: 240–8,000 px.
Aspect ratio : ≤ 8:1
string
Use the provided image URL as the last frame.
The last_frame_url must contain the URL of an uploaded image. The value must be an uploaded file URL, not the file content itself.
Note: Cannot be provided together with reference_*_urlsSupported image types:
  • image/jpeg
  • image/png
  • image/webp
  • image/bmp
Maximum file size: 20 MB
Each side: 240–8,000 px.
Aspect ratio : ≤ 8:1
string[]
Reference videos for the all-purpose reference mode.
The reference_video_urls must contain the URLs of an uploaded videos. The value must be an uploaded file URL, not the file content itself.
Note: The input video duration + duration must not exceed 30 seconds.Supported videos types:
  • .mp4
  • .mov
Supported duration each an clip: 1 to 15 seconds.Maximum reference videos: 5
Maximum each file size: 100 MB
string[]
Reference images for the all-purpose reference mode.
The reference_image_urls must contain the URLs of an uploaded images. The value must be an uploaded file URL, not the file content itself.
Note: Cannot be provided together with the first-frame/last-frame parametersSupported image types:
  • image/jpeg
  • image/png
  • image/webp
  • image/bmp
Maximum reference images: 10
Maximum file size: 20 MB
Aspect ratio : ≤ 8:1
string[]
Reference audio for the all-purpose reference mode.
The reference_audio_urls must contain the URLs of an uploaded audios. The value must be an uploaded file URL, not the file content itself.
Note: The source recommends pairing audio with an image or video rather than using audio alone.Supported audios types:
  • .wav
  • .mp3
Supported duration each an clip: 1 to 15 seconds.Maximum reference audios: 5
Maximum each file size: 15 MB
string[]
File-to-video generation.
The reference_file_urls must contain the URLs of an uploaded file. The value must be an uploaded file URL, not the file content itself.
Note: reference_file_urls cannot be combined with reference_link_urls or the first-frame/last-frame parameters.
  • Maximum 1 file.
  • Maximum size: 100 MB.
  • Supported formats: DOCX, DOC, XLSX, XLS, PPTX, PPT, PDF, TXT, KEY, PAGES, NUMBERS, MD.
  • PDF, DOCX, PPT, KEY, PAGES, and similar documents: up to 50 pages.
Link-to-video generation.
The reference_link_urls must be publicly accessible webpage that does not require login.
Note: reference_link_urls cannot be combined with reference_file_urls or the first-frame/last-frame parameters.Maximum file: 1
string
default:"1080P"
Specify the resolution of the output video.Supported output resolutions:
  • 480P
  • 720P
  • 1080P
Default: 1080P
string
default:"adaptive"
Applies to multi-image generation. In single-image mode, video dimensions follow the source image dimensions.Supported values: 3:4, 4:3, 1:1, 16:9, 9:16, adaptive.Default: adaptive
number
Generated video duration in seconds.Type: integer
Supported range without video input: 230, in one-second steps.
Supported range with reference videos: input video duration + output duration ≤ 30, in one-second steps.
Default: 5
boolean
default:"true"
When enabled, the output video includes an audio track.Default: true
number
The seed can be used to reproduce results.
  • Minimum: 0
  • Maximum: 2147483647
  • If omitted, a random seed is used.
Note: The source schema requires model and input at the top level. It describes prompt as required for text-to-video, but does not mark it as required in the input schema.

Successful Response

number
Response status code.
string
Response message. Contains the error description when the request fails.Example: success
object
required
The task data object containing task id.
string
required
The unique identifier for this task.Example: task_123456

Query Task Status

After submitting a task, use the unified query endpoint to check the task progress and retrieve the generated results.

Get Task Details

Check task status, monitor generation progress, and retrieve results.

Error Response

Response Codes