saurav-samantray/ollama-agent-with-mcp

MCP Server and Ollama AI Agent

This project demonstrates a powerful AI pattern: an autonomous agent that uses external tools to answer questions. It consists of two applications:

1. MCP Server: A FastAPI server that provides a set of tools (Weather, Stocks, Currency, Calculator) via a REST API.
2. AI Agent: A command-line agent powered by a local Ollama LLM that can intelligently decide which tool to use to respond to a user's query.

Prerequisites

1. Python: You must have Python 3.8+ installed.
2. Ollama: You must have Ollama installed and running on your machine.
3. Ollama Model: You need to have a model pulled. We recommend llama3. You can get it by running:

   ollama run llama3
   

   Once you see the >>> prompt, you can type /bye to exit. The model is now available for the agent to use.

Step 1: Set Up and Run the MCP Server

The server provides the tools for the agent.

1. Navigate to the server directory:

   cd mcp_server
   

2. Create a virtual environment and install dependencies:

   python -m venv venv
   source venv/bin/activate  # On Windows, use `venv\Scripts\activate`
   pip install -r requirements.txt
   

3. Run the server:

   python server.py
   

You should see output indicating the server is running on http://127.0.0.1:8000. Keep this terminal window open. You can verify it's working by opening http://127.0.0.1:8000/docs in your browser.

Step 2: Set Up and Run the AI Agent

The agent will connect to the server you just started.

1. Open a new terminal window.

2. Navigate to the agent directory:

   cd ai_agent
   

3. Create a virtual environment and install dependencies:

   python -m venv venv
   source venv/bin/activate  # On Windows, use `venv\Scripts\activate`
   pip install -r requirements.txt
   

4. Run the agent:

   python agent.py
   

Step 3: Interact with the Agent

Once the agent is running, you will see a > prompt. You can now ask it questions that require it to use its tools. The agent will print its "thought process" in real-time.

Example Questions:

• What is the weather like in London? (Uses the get_weather tool)
• What is the price of Apple stock? (Uses the get_stock_price tool)
• How many US dollars is 100 Euros? (This requires two steps: first getting the exchange rate, then using the calculator)
• What is 55 times 3.14? (Uses the calculator tool)
• What is the weather in Tokyo and what is the price of Microsoft stock? (Uses multiple tools)
• Who is the president of the United States? (The agent should answer this from its own knowledge, without using a tool)