SampleMCP

A simple Python project demonstrating how to build and connect MCP servers using FastMCP, LangChain MCP adapters, LangGraph, and OpenAI.

Features

• Math MCP server using stdio
• Weather MCP server using streamable-http
• LangChain MCP client connecting to multiple MCP servers
• LangGraph ReAct agent using MCP tools
• Example tool calls for math and weather queries

Project Structure

SampleMCP/
├── client.py          # MCP client + LangGraph agent
├── mathserver.py      # Math MCP server with add/multiply tools
├── weather.py         # Weather MCP server using streamable HTTP
├── main.py            # Basic sample entry point
├── pyproject.toml     # Project metadata and dependencies
├── requirements.txt   # Python dependencies
└── README.md

Requirements

• Python 3.12+
• OpenAI API key
• Optional: Groq API key

Installation

Clone the repository:

git clone https://github.com/JayantPrakash/SampleMCP.git
cd SampleMCP

Install dependencies:

pip install -r requirements.txt

Or using uv:

uv sync

Environment Variables

Create a .env file in the project root:

OPENAI_API_KEY=your_openai_api_key
GROQ_API_KEY=your_groq_api_key

Running the MCP Servers

1. Start the Weather Server

The weather server uses streamable-http.

python weather.py

By default, it exposes the MCP endpoint at:

http://localhost:8000/mcp

2. Math Server

The math server uses stdio and is started automatically by the client through:

"command": "python",
"args": ["mathserver.py"],
"transport": "stdio"

Running the Client

In a separate terminal, run:

python client.py

The client connects to:

• mathserver.py through stdio
• weather.py through streamable_http

It then creates a LangGraph ReAct agent and asks:

what's (3 + 5) x 12?

and:

what is the weather in California?

MCP Tools

Math Server

Defined in mathserver.py.

add(a: int, b: int) -> int

Adds two numbers.

multiple(a: int, b: int) -> int

Multiplies two numbers.

Weather Server

Defined in weather.py.

get_weather(location: str) -> str

Returns a sample weather response for a given location.

Notes

• mathserver.py uses transport="stdio".
• weather.py uses transport="streamable-http".
• In client.py, the streamable HTTP transport is configured as streamable_http.
• Make sure the weather server is running before executing client.py.

Example Output

Math response: 96
Weather response: It's always raining in California