ocr-server

OCR Server

A lightweight and async OCR server powered by FastAPI, Celery, EasyOCR, and Redis. Supports image and PDF OCR using CPU.

---

Features

• OCR for images (.jpg, .png, etc.) and PDFs
• Async task processing with Celery
• Accurate OCR using EasyOCR
• Task-based API (submit → get result)
• Modular codebase (production-ready)

---

📁 Project Structure

ocr-server/
├── Dockerfile
├── docker-compose.yml
├── requirements.txt
├── src/
│   ├── __main__.py       # FastAPI app entrypoint
│   ├── worker.py         # Celery worker app
│   ├── tasks.py          # Celery OCR task
│   ├── ocr_engine.py     # EasyOCR logic
│   ├── utils.py          # PDF/Image tools
│   └── config.py         # Paths & settings

---

⚙️ Requirements

• Docker & Docker Compose

---

🐳 Run with Docker

docker-compose up --build

---

📡 API Endpoints

Submit OCR task

POST /ocr
Form-Data: file=@yourfile.pdf

Get result

GET /result/{task_id}

---

✅ Example Response

{
  "status": "completed",
  "result": {
    "type": "pdf",
    "pages": [
      {
        "page": 1,
        "text": [
          {
            "text": "Hello World",
            "confidence": 0.98,
            "box": [[x1, y1], [x2, y2], ...]
          }
        ]
      }
    ]
  }
}

---

🧪 Test with curl

curl -F "file=@sample.pdf" http://localhost:8000/ocr
curl http://localhost:8000/result/<task_id>

This site is open source. Improve this page.