Skip to content

PixVerse V6 Fusion / Reference-to-Video ​

This page is generated from the upstream official OpenAPI documentation. Endpoints, parameters and examples match the official docs; brand and API domain are replaced with ModelFold.

Endpoint ​

POST /jobs/createTask

Full URL: https://ai.modelfold.online/v1/jobs/createTask

Authentication: every request must include the Authorization: Bearer <API_KEY> header.

Query Task Status ​

After submitting a task, you can check its progress and retrieve the result through the unified query endpoint:

  • Get Task Details — Learn how to check task status and retrieve the generated result

TIP

In production, we recommend using the callBackUrl parameter to receive an automatic notification when generation is complete instead of polling the status endpoint.

TIP

Enabling audio costs more. For current prices, see ai.modelfold.online/pricing.

Request Parameters ​

Body (application/json) ​

ParameterTypeRequiredDescription
modelstringYesThe name of the model used for the generation task. Required field. Default:"pixverse-v6/reference-to-video"。 Allowed values:pixverse-v6/reference-to-video。 示例:pixverse-v6/reference-to-video。
callBackUrlstringNoThe callback URL to receive the completion notification of the generation task. Optional configuration, recommended for production environments.
- After the task generation is completed, the system will push the task status and results to this URL via POST
- The callback content includes the URL of the generated content and task-related information
- Your callback interface must support receiving POST requests and JSON-formatted request bodies
- You can also actively poll the task status by calling the task details interface 示例:https://your-domain.com/api/callback。
inputobjectYesInput parameters for the generation task
  input.promptstringYesGenerate prompt, cannot be empty, length is limited to 3-5000 characters. 示例:A cinematic sunrise illuminates a mist-shrouded mountain lake; the camera slowly sweeps across the water's surface as a flock of birds flies overhead.。
  input.image_referencesarray&lt;object&gt;YesFusion multi-reference image list.
ref_name must be unique within the same list. You can reference the reference object in the prompt via @ref_name.
    input.image_references[].image_urlstringYesReference image URL. 示例:https://your-domain.com/images/reference.png。
    input.image_references[].typestringNoReference image type. subject means subject, background means background. Default:"subject"。 Allowed values:subject / background。 示例:subject。
    input.image_references[].ref_namestringNoReference image name, maximum 30 characters. Can be referenced in the prompt using @name. 示例:dog。
  input.aspect_ratiostringYesRequired. Output video aspect ratio. Supported in text-to-video and Fusion multi-reference image to video modes. Allowed values:16:9 / 4:3 / 1:1 / 3:4 / 9:16 / 2:3 / 3:2 / 21:9。 示例:16:9。
  input.qualitystringYesRequired. Output video resolution. Supports 360p, 540p, 720p, and 1080p. Allowed values:360p / 540p / 720p / 1080p。 示例:720p。
  input.durationintegerYesRequired. Output video duration in seconds, 1–15. 示例:5。
  input.generate_audio_switchbooleanNoWhether to generate audio synchronized with the video content. Default:false。 示例:false。
  input.seedintegerNoRandom seed, value range is 0-2147483647. Using the same parameters and seed helps improve the reproducibility of the results. 示例:123456789。

Request Example ​

bash
curl -X POST "https://ai.modelfold.online/v1/jobs/createTask" \
  -H "Authorization: Bearer $MODELFOLD_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "model": "pixverse-v6/reference-to-video",
  "input": {
    "prompt": "Use the reference image identity and generate a cinematic fashion video",
    "image_references": [
      {
        "image_url": "https://example.com/reference-1.png"
      },
      {
        "image_url": "https://example.com/reference-2.png"
      }
    ],
    "duration": 5,
    "quality": "720p",
    "aspect_ratio": "9:16",
    "generate_audio_switch": false,
    "seed": 123456
  },
  "callBackUrl": "https://example.com/ModelFold/callback"
}'
json
{
  "model": "pixverse-v6/reference-to-video",
  "input": {
    "prompt": "Use the reference image identity and generate a cinematic fashion video",
    "image_references": [
      {
        "image_url": "https://example.com/reference-1.png"
      },
      {
        "image_url": "https://example.com/reference-2.png"
      }
    ],
    "duration": 5,
    "quality": "720p",
    "aspect_ratio": "9:16",
    "generate_audio_switch": false,
    "seed": 123456
  },
  "callBackUrl": "https://example.com/ModelFold/callback"
}

Response ​

200 — Request successful. ​

Callback ​

videoTaskCompleted ​

Callback URL: {request.body#/callBackUrl} (the callBackUrl in the request body).

The system sends this callback when the pixverse-v6/reference-to-video task succeeds or fails.

Callback body parameters

参数类型必填说明
codeinteger是Unified callback code: 200 for success, 501 for failure. Allowed values:200 / 501。 示例:200。
msgstring是Unified callback message. Allowed values:Playground task completed successfully. / Playground task failed.。 示例:Playground task completed successfully.。
dataobject是—
  data.taskIdstring是The unique task identifier.
  data.modelstring是The model used for the task. Allowed values:pixverse-v6/reference-to-video。 示例:pixverse-v6/reference-to-video。
  data.statestring是Terminal task state. Allowed values:success / fail。 示例:success。
  data.paramstring是JSON string containing the submitted task parameters.
  data.resultJsonstring否JSON string containing resultUrls, firstFrameUrl, lastFrameUrl, or resultObject when successful. Present only when state is success.
  data.failCodestring否Failure code, a numeric string such as "400" or "500". Present only when state is fail.
  data.failMsgstring否Failure reason. Present only when state is fail.
  data.costTimeinteger否Processing time in seconds. Present on both success and failure callbacks.
  data.completeTimeinteger否Task completion timestamp.
  data.createTimeinteger否Task creation timestamp.
  data.updateTimeinteger否Use completeTime as the completion time; updateTime may equal createTime.
  data.creditsConsumednumber否Credits consumed by the task; 0 for failed tasks.

Task completed successfully

json
{
  "code": 200,
  "msg": "Playground task completed successfully.",
  "data": {
    "taskId": "task_example_video_webhook_001",
    "model": "pixverse-v6/reference-to-video",
    "state": "success",
    "param": "{\"input\":\"{}\",\"callBackUrl\":\"https://example.com/callback\",\"model\":\"pixverse-v6/reference-to-video\"}",
    "resultJson": "{\"resultUrls\":[\"https://example.com/generated-video.mp4\"]}",
    "costTime": 12,
    "completeTime": 1234567890,
    "createTime": 1234560000,
    "updateTime": 1234567890,
    "creditsConsumed": 1.23
  }
}

Task failed

json
{
  "code": 501,
  "msg": "Playground task failed.",
  "data": {
    "taskId": "task_example_video_webhook_001",
    "model": "pixverse-v6/reference-to-video",
    "state": "fail",
    "param": "{\"input\":\"{}\",\"callBackUrl\":\"https://example.com/callback\",\"model\":\"pixverse-v6/reference-to-video\"}",
    "failCode": "400",
    "failMsg": "<error detail>",
    "costTime": 12,
    "creditsConsumed": 0
  }
}

ModelFold — one API for all the best AI models.