API Reference

Async Tasks API

Image, video, and other high-latency models use a unified task interface. Submitting a task returns its status; video tasks require polling, while some image tasks return the result immediately.

Submit

POST /v1/tasks

Create a task; video returns running, some images return succeeded immediately.

Query

GET /v1/tasks/{id}

Retrieve task status, progress, output, and errors.

List

GET /v1/tasks

List the most recent tasks for the current API key user.

Request Parameters

modelRequiredImage or video model ID, e.g. wan2.6-t2i, wan2.6-i2v, pixverse-v6, happyhorse-1.0-t2v.

promptRequiredGeneration prompt. For video, describe the subject, motion, camera, scene, and lighting.

sizeOptionalImage or video dimensions, e.g. 1024x1024, 1280x720.

durationVideo OptionalVideo duration, choose 5, 8, 10, 15, etc. based on model capabilities.

img_urlImage-to-Video OptionalReference image URL, used for i2v/r2v type tasks.

negative_promptOptionalElements, styles, or actions you don't want to appear.

webhook_urlNot currently supportedThe current public task endpoint does not trigger webhooks; production systems should use GET /v1/tasks/{id} polling.

Submit Task

curl https://nexusflow.vip/v1/tasks \
  -H "Authorization: Bearer $API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "wan2.6-t2i",
    "prompt": "Generate a tech product poster",
    "size": "1024x1024"
  }'

Query Task

curl https://nexusflow.vip/v1/tasks/task_xxx \
  -H "Authorization: Bearer $API_KEY"

View Full Tasks Documentation