MCP Agent - AI Expense Tracker

# Expense Tracker API A simple FastAPI application for tracking expenses with SQLite database and MCP server integration. ## Features - **Web Dashboard**: Beautiful, modern UI for viewing expense analytics - **SQLite Database**: Persistent storage for transactions - **REST API**: Full CRUD operations for expense tracking - **MCP Server**: AI assistant integration via Model Context Protocol - **Simple & Clean**: Minimal dependencies, easy to understand ## Quick Start 1. **Install dependencies**: ```bash pip install -r requirements.txt ``` 2. **Run both API and MCP servers**: ```bash python start.py ``` Or run separately: ```bash # API server python main.py # MCP server (in another terminal) python mcp_server.py ``` 3. **Access the application**: - **Dashboard**: http://localhost:8002/ (Main web interface) - **API Docs**: http://localhost:8002/docs (Interactive API documentation) - **MCP Server**: http://localhost:9002 (AI assistant integration) 4. **Populate sample data** (optional): ```bash python seed_data.py ``` ## Project Structure ``` server/ ├── main.py # FastAPI application with routes ├── models.py # Pydantic models ├── store.py # SQLite data store ├── config.py # Configuration settings ├── mcp_server.py # MCP server for AI integration ├── start.py # Startup script ├── seed_data.py # Sample data populator ├── templates/ │ └── index.html # Dashboard web interface ├── requirements.txt # Dependencies └── expenses.db # SQLite database (auto-created) ``` ## Dashboard Features The web dashboard provides: - **Real-time Statistics**: View total income, expenses, and net balance at a glance - **Category Breakdown**: Visual summary showing spending by category with percentages - **Latest Transactions**: Quick view of recent financial activities - **Auto-refresh**: Dashboard updates every 30 seconds automatically - **Responsive Design**: Works beautifully on desktop and mobile devices - **Dark Theme**: Easy on the eyes with modern dark mode UI Access the dashboard at: **http://localhost:8002/** ## API Endpoints ### Dashboard - `GET /` - Web dashboard interface (HTML) ### Transactions - `GET /transactions` - List all transactions - `GET /transactions/{id}` - Get specific transaction - `POST /transactions` - Create new transaction - `PUT /transactions/{id}` - Update transaction - `DELETE /transactions/{id}` - Delete transaction - `GET /transactions/search?q={query}` - Search transactions ### Analytics - `GET /summary` - Get financial summary - `GET /summary/categories` - Category breakdown - `GET /transactions/type/{type}` - Get by type (income/expense) - `GET /transactions/category/{category}` - Get by category - `GET /transactions/date-range` - Get by date range ### System - `GET /health` - Health check ## Database Uses SQLite for persistent storage. Database file: `expenses.db` ## MCP Integration The MCP server exposes all API endpoints as tools for AI assistants. Connect your agent to: ``` http://localhost:9002/mcp ``` ## Configuration Environment variables (optional): - `HOST` - Server host (default: 0.0.0.0) - `PORT` - API port (default: 8002) - `MCP_HOST` - MCP host (default: localhost) - `MCP_PORT` - MCP port (default: 9002)