PaddleOCR Test

API Usage

POST an image to the OCR endpoint. The server accepts one multipart file field named file and returns merged text plus per-line details.

curl -X POST https://ocr.leejh.cyou/ocr \
  -F "file=@your_image.png"

Use Python requests library.

import requests

with open("your_image.png", "rb") as f:
    r = requests.post(
        "https://ocr.leejh.cyou/ocr",
        files={"file": f}
    )
print(r.json())

Browser or Node.js with fetch.

const form = new FormData();
form.append("file", fileInput.files[0]);

const res = await fetch("https://ocr.leejh.cyou/ocr", {
  method: "POST",
  body: form
});
console.log(await res.json());

Accepts one required multipart file field named file. The server currently validates Content-Type with the rule image/*, so common image formats supported by clients and OCR pipelines such as PNG, JPG/JPEG, BMP, WEBP, GIF, and TIFF can be uploaded as long as the request is labeled as an image.

Size limit: 10MB per image. If no text is recognized, the API still returns 200 OK with {"text": "", "lines": []}.

Request Example

POST /ocr
Content-Type: multipart/form-data

file=@your_image.png

Response Example

{
  "text": "示例识别文本",
  "lines": [
    {
      "text": "示例识别文本",
      "confidence": 0.9987,
      "box": [[12.5, 18.0], [220.1, 18.0], [220.1, 52.3], [12.5, 52.3]]
    }
  ]
}

Field Reference

file                required image file field (multipart/form-data)
text                concatenated OCR text from the whole image
lines               array of recognized text lines
lines[].text        text content of the line
lines[].confidence  confidence score, range 0-1
lines[].box         quadrilateral coordinates [[x1,y1],[x2,y2],[x3,y3],[x4,y4]]

Error Response Examples

400 Bad Request
{"detail": "Only image files are accepted"}

400 Bad Request
{"detail": "Image too large (max 10MB)"}

500 Internal Server Error
{"detail": ""}

Service URL: https://ocr.leejh.cyou · Health check: GET /health

PaddleOCR Test

Upload Image

Result

API Usage