POST /jobs/createTask

POST /api/v1/jobs/createTask queues an asynchronous AI generation job and returns a taskId. HTTP 200 means the task was accepted into the queue, not that generation finished. Poll GET /jobs/recordInfo?taskId=… until state is success or fail. Credits are validated before accept; deduction happens only on success.

Creates an async generation task.

Request

{
  "model": "google/veo-3-1",
  "input": {
    "prompt": "A cinematic drone shot over mountains at sunrise",
    "aspect_ratio": "16:9",
    "duration": "5"
  }
}

Response

{
  "code": 200,
  "msg": "success",
  "data": { "taskId": "task_xxx" }
}

Notes

• Image vs video routing is inferred from the model id when capability is omitted
• Optional capability: "image" | "video"
• Your account must have enough credits for the quoted model before the task is accepted (HTTP 402 if not)
• Credits are not deducted when taskId is returned — billing happens when recordInfo reports state=success