WuyinKeji GPT-Image-2 Image Generation

This skill wraps the WuyinKeji (速创科技) GPT-Image-2 async image generation API.

API Endpoints

Authentication

• Header: Authorization: <api_key>
• Body: also include "key": "<api_key>" in JSON payload

1. Submit Generation Task

Request

curl -s -X POST "https://api.wuyinkeji.com/api/async/image_gpt" \
  -H "Content-Type: application/json" \
  -H "Authorization: <API_KEY>" \
  -d '{
    "key": "<API_KEY>",
    "prompt": "Your image generation prompt here",
    "size": "16:9",
    "urls": ["http://example.com/ref.jpg"]
  }'

Parameters

Response

{
  "code": 200,
  "msg": "成功",
  "data": {
    "id": "image_xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
    "count": "1"
  }
}

Save the data.id — this is the async task ID.

2. Query Result (Polling)

curl -s "https://api.wuyinkeji.com/api/async/detail?id=<TASK_ID>&key=<API_KEY>"

Response Fields

Example (completed)

{
  "code": 200,
  "data": {
    "status": 2,
    "result": ["https://openpt1.wuyinkeji.com/xxxx.png"],
    "created_at": "2026-04-22 13:42:01",
    "updated_at": "2026-04-22 13:43:17"
  }
}

3. Download Image

curl -sL "<RESULT_URL>" -o output.png

Typical resolution: 1672×941 (the API output is fixed at this resolution regardless of the size parameter; use the size parameter mainly for aspect ratio hints).

Model Capabilities

• Text rendering: Good Chinese/English text support in images (prompt explicitly for sharp, readable typography)
• Reference images: Supports up to 14 reference URLs for style/person consistency
• Realism: Improved skin texture, hair, facial details over earlier DALL·E models
• Color accuracy: Reduced yellow/orange color cast compared to DALL·E

Tips for Quality