Nutrient Document Engine MCP Server

# Procurement Agent TypeScript AI-powered procurement document processing system using LangGraph.js and Document Engine MCP Server. ## Overview This system automatically processes procurement documents and performs 3-way matching between: - **Purchase Orders (PO)** - **Invoices** - **Payments** It uses AI agents to classify documents, extract data, find matches, and generate comprehensive reports. ## Prerequisites - Node.js 18+ - Document Engine running locally - MCP server running (default: http://localhost:5100/mcp) - OpenAI API key set in the environment ## Setup 1. **Install dependencies**: ```bash pnpm install ``` 2. **Environment setup**: Create a `.env` file in this directory with: ```env OPENAI_API_KEY=your_openai_api_key_here MCP_SERVER_URL=http://localhost:5100/mcp ``` 3. **Start Document Engine MCP Server**: From the main project directory: ```bash pnpm build && MCP_TRANSPORT=http pnpm start ``` 4. **Run the example**: ```bash pnpm start ``` The system will automatically: 1. Load all PDFs from `../../langgraphjs/assets/procurement/pdfs/` 2. Upload them to Document Engine 3. Classify each document using AI 4. Extract structured data (PO numbers, invoice numbers, amounts, etc.) 5. Perform 3-way matching to find related documents 6. Generate a comprehensive report with insights