mcp-house-assistant

MCP House Assistant

A modular AI-powered assistant that uses OpenAI to interpret natural language queries and route them to the correct tool (market recap or weather).

🚦 Usage Modes

• CLI Tool: Run queries directly from your terminal.
• HTTP API Endpoint: Start a server and send queries via HTTP POST requests.

---

🏃 Quick Start

CLI Tool

Run the CLI and enter your query:

python mcp_server.py "What's the weather in London?"

Or just run:

python mcp_server.py

And enter your query when prompted.

HTTP API Endpoint

1. Start the server:

   uvicorn mcp_api_server:app --reload
   

2. Send a POST request to /query:

   curl -X POST "http://localhost:8000/query" \
     -H "Content-Type: application/json" \
     -d '{"query": "What's the weather in London?"}'
   

   You will receive a JSON response with the result.

---

🚀 How It Works

• Natural Language Interface: Ask questions like "What's the weather in London?" or "Give me a market recap for today".
• LLM Orchestration: Uses OpenAI's GPT-4o to decide which tool to call and with what parameters.
• Tool Execution: Calls the appropriate Python tool and prints the result as JSON.
• No Web Server Needed: Everything runs from the command line. (The old FastAPI server is deprecated and has been removed.)

🛠️ Tools

• Market Recap Tool: Fetches market data for major indices or specified symbols for a given date.
• Weather Tool: Fetches current weather and forecast for a given location using WeatherAPI.

🗝️ Setup

1. Clone the repository and enter the project directory:

   cd "MCP hiouse assistant"
   

2. (Recommended) Create and activate a virtual environment:

   python3 -m venv .venv
   source .venv/bin/activate
   

3. Install dependencies:

   pip install -r requirements.txt
   # Or, if requirements.txt is missing:
   pip install openai python-dotenv yfinance requests
   

4. Set up your .env file:
   Create a .env file in the project root with:

   OPENAI_API_KEY=your_openai_api_key_here
   WEATHER_API_KEY=your_weatherapi_key_here
   

   Replace with your actual API keys.

🧪 Running Tests

Tests are located in the `