README.md

metadata

title: OCR Scan Vision API
emoji: 🧾
colorFrom: indigo
colorTo: blue
sdk: docker
pinned: false
license: mit

Arabic OCR API

Simple API for extracting Arabic text from contract images.

Endpoint

POST /ocr

Request

• file: Image file (multipart/form-data)
• min_conf: Minimum confidence threshold (0.0-1.0, optional)

Response

{
  "items": [
    {"box_id": 1, "text": "النص", "confidence": 0.95, "bbox": [x1,y1,x2,y2]}
  ],
  "text": "النص الكامل",
  "total_boxes": 10
}

Usage

curl -X POST "https://sharshar1-ocr.hf.space/ocr" -F "file=@image.jpg"