📄 POC for using Docling to parse documents to structured output for AI analysis

A proof-of-concept application that combines Docling's powerful document parsing capabilities with OpenAI's API to enable intelligent analysis and interaction with your documents.

🎯 Project Goal

This project demonstrates how to:

1. Parse Documents - Use Docling to extract structured content from various document formats (PDF, DOCX, PPTX, images)
2. Process with AI - Send the parsed content to OpenAI's API along with user-provided prompts
3. Get Intelligent Output - Receive AI-generated responses that can analyze, summarize, answer questions, or perform other tasks on your document content

🚀 How It Works

Document Input → Docling Parser → Structured Content → OpenAI API + User Prompt → AI Response

1. Document Ingestion: Upload or provide a document (PDF, DOCX, PPTX, or image)
2. Content Extraction: Docling processes the document and converts it to structured Markdown/JSON
3. Prompt Integration: Combine the parsed content with your custom text prompt
4. AI Analysis: Send both to OpenAI's API for intelligent processing
5. Response: Receive AI-generated insights, summaries, answers, or analysis

🔧 Features

• Multi-format Support: Handle PDFs, Word documents, PowerPoint presentations, and images
• Structured Output: Convert documents to clean, structured Markdown or JSON
• Custom Prompting: Add your own prompts to guide the AI analysis
• OpenAI Integration: Leverage GPT models for powerful document understanding
• Flexible Pipeline: Easy to extend for different use cases (RAG, Q&A, summarization, etc.)

🛠️ Tech Stack

• Docling - Advanced document parsing and conversion
• OpenAI API - GPT models for intelligent text processing
• Python - Core application framework

📝 Example Workflow

# 1. Parse document with Docling
parsed_content = docling.parse("document.pdf")

# 2. Combine with user prompt
user_prompt = "Summarize the key findings in this research paper"
full_prompt = f"{user_prompt}\n\nDocument content:\n{parsed_content}"

# 3. Send to OpenAI
response = openai.chat.completions.create(
    model="gpt-4",
    messages=[{"role": "user", "content": full_prompt}]
)

# 4. Get AI analysis
ai_response = response.choices[0].message.content

🚀 Quick Start

Prerequisites

• Python 3.8+
• Node.js 16+
• OpenAI API key

Backend Setup

1. Clone the repository

   git clone <your-repo-url>
   cd doc-parse-poc

2. Set up Python environment

   cd backend
   python -m venv venv
   source venv/bin/activate  # On Windows: venv\Scripts\activate
   pip install -r requirements.txt

3. Configure environment variables

   cp .env.example .env
   # Edit .env and add your OpenAI API key

4. Run the backend

   python app.py

   The API will be available at http://localhost:5000

Frontend Setup

1. Install dependencies

   cd frontend
   npm install

2. Start the development server

   npm start

   The app will open at http://localhost:3000

📖 Usage

1. Upload a Document: Drag and drop or click to select a document (PDF, DOCX, PPTX, or image)
2. Enter a Prompt: Type what you want to know about the document
3. Analyze: Click the analyze button to process your document
4. View Results: See the AI-generated analysis with document metadata

Example Prompts

• "Summarize the key points and main conclusions"
• "What are the most important findings?"
• "Extract all numerical data and statistics"
• "Identify the main themes and topics"
• "What action items or recommendations are mentioned?"

🏗️ Architecture

Backend (Python/Flask)

backend/
├── app.py                 # Main Flask application
├── config/
│   └── settings.py        # Configuration management
├── services/
│   ├── document_parser.py # Docling integration
│   ├── openai_service.py  # OpenAI API client
│   └── file_service.py    # File handling
├── routes/
│   ├── document_routes.py # Document processing endpoints
│   └── health_routes.py   # Health check endpoints
├── requirements.txt       # Python dependencies
└── .env.example          # Environment template

Frontend (React/TypeScript)

frontend/
├── src/
│   ├── components/        # React components
│   ├── services/          # API client
│   ├── types/            # TypeScript definitions
│   └── utils/            # Helper functions
├── public/               # Static assets
└── package.json         # Node.js dependencies

🔧 Configuration

Environment Variables

Backend (.env)

OPENAI_API_KEY=your_api_key_here
OPENAI_MODEL=gpt-4
FLASK_DEBUG=False
SECRET_KEY=your_secret_key
UPLOAD_FOLDER=uploads
MAX_CONTENT_LENGTH=16777216  # 16MB

Frontend

REACT_APP_API_URL=http://localhost:5000/api

🚀 Production Deployment

Backend (Flask)

# Install production server
pip install gunicorn

# Run with gunicorn
gunicorn --bind 0.0.0.0:5000 wsgi:application

Frontend (React)

# Build for production
npm run build

🧪 API Reference

POST /api/analyze

Analyze a document with AI

Request:

• file: Document file (multipart/form-data)
• prompt: Analysis prompt (string)

Response:

{
	"success": true,
	"analysis": "AI-generated analysis...",
	"metadata": {
		"document": {
			"title": "document.pdf",
			"file_type": ".pdf",
			"file_size": 1024,
			"page_count": 5
		},
		"usage": {
			"total_tokens": 1500,
			"model_used": "gpt-4"
		}
	}
}

GET /api/supported-formats

Get supported file formats

GET /api/health

Health check endpoint

🤝 Contributing

1. Fork the repository
2. Create a feature branch
3. Make your changes
4. Add tests if applicable
5. Submit a pull request

📄 License

MIT License (inherits core Docling license)

---

Ready to unlock the intelligence hidden in your documents? This POC shows you how to combine state-of-the-art document parsing with powerful AI analysis!