> ## Documentation Index
> Fetch the complete documentation index at: https://docs.yingtu.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# YingTu API

> One asynchronous task flow for image and video generation.

YingTu exposes a single task-shaped workflow for generative media. Submit a
request, store the returned task ID, then receive the result by polling or by
callback. Switching models changes the `model` and `input` fields—not the
overall lifecycle.

<Info>
  This first-stage reference defines the unified task contract and documents
  Nano Banana first. Additional model pages are scaffolded and remain clearly
  marked until their public contracts are verified.
</Info>

## One lifecycle, model-specific inputs

Every submission uses the same envelope:

```json theme={null}
{
  "model": "gemini-3.1-flash-image",
  "input": {
    "prompt": "A quiet reading room at sunrise",
    "aspect_ratio": "16:9",
    "resolution": "2K"
  },
  "callback_url": "https://example.com/webhooks/yingtu"
}
```

The model pages focus on what changes: supported inputs, constraints, outputs,
and examples. The Unified API section describes what stays the same.

## Where to go next

<CardGroup cols={2}>
  <Card title="Quickstart" icon="rocket" href="/en/getting-started/quickstart">
    Submit a task and retrieve its output.
  </Card>

  <Card title="Async task API" icon="clock-3" href="/en/unified-api/async-tasks">
    Understand states, polling, callbacks, and retries.
  </Card>

  <Card title="Nano Banana" icon="image" href="/en/models/image/nano-banana/overview">
    Start with the first documented image model family.
  </Card>

  <Card title="API Playground" icon="terminal-square" href="/en/api-reference/submit-task">
    Inspect the OpenAPI schema and prepare a request.
  </Card>
</CardGroup>