API

Every alibaba.api.* operation is listed below with parameter shapes and return types from the plugin Zod schemas.

New to Corsair? See API access, authentication, and error handling.

Image

generate

image.generate Generate or edit images with Qwen-Image 3.0 or Wan 2.7 Image; returns image URLs valid for 24 hours Risk: write

await corsair.alibaba.api.image.generate({});

Input

Name	Type	Required	Description
`model`	`qwen-image-3.0-pro \| qwen-image-3.0 \| wan2.7-image-pro \| wan2.7-image`	Yes	—
`prompt`	`string`	Yes	—
`images`	`string[]`	No	—
`parameters`	`object`	No	—

parameters full type

{
  size?: string,
  n?: number,
  seed?: number,
  watermark?: boolean,
  negative_prompt?: string,
  prompt_extend?: boolean,
  prompt_extend_mode?: direct | agent,
  enable_thinking?: boolean,
  thinking_mode?: boolean,
  enable_sequential?: boolean,
  bbox_list?: tuple[][],
  color_palette?: {
    hex: string,
    ratio: string
  }[]
}

Output

Name	Type	Required	Description
`output`	`object`	Yes	—
`usage`	`object`	No	—
`request_id`	`string`	Yes	—

output full type

{
  rewrite_status?: string | null,
  choices: {
    finish_reason?: string | null,
    message: {
      role: string,
      content: {
        image: string
      }[]
    }
  }[]
}

usage full type

{
  output_width?: number | null,
  output_height?: number | null,
  input_image_count?: number | null,
  output_image_count?: number | null,
  input_image_type?: string | null,
  output_image_type?: string | null,
  width?: number | null,
  height?: number | null,
  image_count?: number | null
}

Tasks

get

tasks.get Get a generation task status and, once it succeeds, the video URL (valid 24 hours) Risk: read

await corsair.alibaba.api.tasks.get({});

Input

Name	Type	Required	Description
`task_id`	`string`	Yes	—

Output

Name	Type	Required	Description
`output`	`object`	Yes	—
`usage`	`object`	No	—
`request_id`	`string`	Yes	—

output full type

{
  task_id: string,
  task_status: PENDING | RUNNING | SUCCEEDED | FAILED | CANCELED | UNKNOWN,
  submit_time?: string | null,
  scheduled_time?: string | null,
  end_time?: string | null,
  video_url?: string | null,
  orig_prompt?: string | null,
  code?: string | null,
  message?: string | null
}

usage full type

{
  duration?: number | null,
  input_video_duration?: number | null,
  output_video_duration?: number | null,
  SR?: number | null,
  ratio?: string | null,
  video_count?: number | null
}

Video

edit

video.edit Start a Wan 2.7 video editing task on an input video Risk: write

await corsair.alibaba.api.video.edit({});

Input

Name	Type	Required	Description
`model`	`wan2.7-videoedit`	Yes	—
`prompt`	`string`	No	—
`negative_prompt`	`string`	No	—
`media`	`object[]`	Yes	—
`parameters`	`object`	No	—

media full type

{
  type: video | reference_image,
  url: string | string
}[]

parameters full type

{
  resolution?: 720P | 1080P,
  ratio?: 16:9 | 9:16 | 1:1 | 4:3 | 3:4,
  duration?: 0 | number,
  audio_setting?: auto | origin,
  prompt_extend?: boolean,
  watermark?: boolean
}

Output

Name	Type	Required	Description
`output`	`object`	Yes	—
`request_id`	`string`	Yes	—

output full type

{
  task_id: string,
  task_status: PENDING | RUNNING | SUCCEEDED | FAILED | CANCELED | UNKNOWN
}

imageToVideo

video.imageToVideo Start a Wan 2.7 image-to-video task from a first frame, first and last frames, or a clip to continue Risk: write

await corsair.alibaba.api.video.imageToVideo({});

Input

Name	Type	Required	Description
`model`	`wan2.7-i2v`	Yes	—
`prompt`	`string`	No	—
`negative_prompt`	`string`	No	—
`media`	`object[]`	Yes	—
`parameters`	`object`	No	—

media full type

{
  type: first_frame | last_frame | driving_audio | first_clip,
  url: string
}[]

parameters full type

{
  resolution?: 720P | 1080P,
  duration?: number,
  prompt_extend?: boolean,
  watermark?: boolean,
  seed?: number
}

Output

Name	Type	Required	Description
`output`	`object`	Yes	—
`request_id`	`string`	Yes	—

output full type

{
  task_id: string,
  task_status: PENDING | RUNNING | SUCCEEDED | FAILED | CANCELED | UNKNOWN
}

referenceToVideo

video.referenceToVideo Start a Wan 2.7 reference-to-video task from reference images, videos, and voices Risk: write

await corsair.alibaba.api.video.referenceToVideo({});

Input

Name	Type	Required	Description
`model`	`wan2.7-r2v`	Yes	—
`prompt`	`string`	Yes	—
`negative_prompt`	`string`	No	—
`media`	`object[]`	Yes	—
`parameters`	`object`	No	—

media full type

{
  type: reference_image | reference_video | first_frame,
  url: string,
  reference_voice?: string
}[]

parameters full type

{
  resolution?: 720P | 1080P,
  ratio?: 16:9 | 9:16 | 1:1 | 4:3 | 3:4,
  duration?: number,
  prompt_extend?: boolean,
  watermark?: boolean,
  seed?: number
}

Output

Name	Type	Required	Description
`output`	`object`	Yes	—
`request_id`	`string`	Yes	—

output full type

{
  task_id: string,
  task_status: PENDING | RUNNING | SUCCEEDED | FAILED | CANCELED | UNKNOWN
}

textToVideo

video.textToVideo Start a Wan text-to-video task (Wan 3.0 also takes reference media, files, and links); returns a task id Risk: write

await corsair.alibaba.api.video.textToVideo({});

Input

Name	Type	Required	Description
`model`	`wan2.7-t2v \| wan3.0-video \| wan3.0-video-prime`	Yes	—
`prompt`	`string`	No	—
`negative_prompt`	`string`	No	—
`audio_url`	`string`	No	—
`media`	`object[]`	No	—
`parameters`	`object`	No	—

media full type

{
  type: first_frame | last_frame | reference_image | reference_video | reference_audio | file | link,
  url: string
}[]

parameters full type

{
  resolution?: 480P | 720P | 1080P,
  ratio?: 16:9 | 9:16 | 1:1 | 4:3 | 3:4 | 21:9 | adaptive,
  duration?: -1 | number,
  audio?: boolean,
  prompt_extend?: boolean,
  watermark?: boolean,
  seed?: number
}

Output

Name	Type	Required	Description
`output`	`object`	Yes	—
`request_id`	`string`	Yes	—

output full type

{
  task_id: string,
  task_status: PENDING | RUNNING | SUCCEEDED | FAILED | CANCELED | UNKNOWN
}

Overview

Plugins

Image

generate

Tasks

get

Video

edit

imageToVideo

referenceToVideo

textToVideo

​Image

​generate

​Tasks

​get

​Video

​edit

​imageToVideo

​referenceToVideo

​textToVideo

Image

generate

Tasks

get

Video

edit

imageToVideo

referenceToVideo

textToVideo