Python MCP Server Client

🚀 Complete Guide to Building MCP Server and MCP Client from Scratch

This is a comprehensive tutorial on how to build a Model Context Protocol (MCP) server and client from scratch using Python.

🚀 Quick Start

✨ Features

• Build MCP Server and Client from scratch using Python.
• Utilize various Python packages for different functionalities.

📦 Installation

Before you start, make sure you have installed the following Python packages:

pip install python-multipart requests asyncio websockets uvicorn serper-dev-api

These packages include:

• python-multipart: Handle multipart requests.
• requests: Send HTTP requests.
• asyncio and websockets: Implement asynchronous communication and WebSocket support.
• uvicorn: Run FastAPI applications.
• serper-dev-api: Use Serper.dev's API to parse web page content.

💻 Usage Examples

Basic Usage

First, initialize the project by creating a new directory and adding the following files:

server.py

This is the core code of the MCP Server, responsible for handling WebSocket connections and tool calls.

import asyncio
import websockets
import json
from typing import List, Dict
from fastapi import FastAPI, HTTPException
from fastapi.responses import JSONResponse
from serper-dev-api import Serper

app = FastAPI()
serper = Serper(your_api_key)  # Fill in your Serper API Key

# Define the tool interface
class Tool:
    def __init__(self, name: str, description: str, input_schema: Dict):
        self.name = name
        self.description = description
        self.input_schema = input_schema

async def serve():
    async with websockets.connect("ws://localhost:8000") as websocket:
        await websocket.accept()
        print("Connection established!")
        
        while True:
            try:
                # Receive tool call requests
                tool_call = await websocket.recv()
                print(f"Received tool call: {tool_call}")
                
                # Parse tool parameters
                params = json.loads(tool_call)
                tool_name = params["name"]
                arguments = params["arguments"]
                
                if tool_name == "search":
                    result = serper.search(arguments, num_results=3)
                    await websocket.send(json.dumps(result))
                elif tool_name == "parse_website":
                    # Assume you have a function to parse web page content
                    content = parse_website(arguments)
                    await websocket.send(content)
                
            except Exception as e:
                print(f"Error: {str(e)}")

client.py

This is the core code of the MCP Client, responsible for communicating with the Server and handling user queries.

import asyncio
from typing import Dict, List
from fastapi import FastAPI, HTTPException
from fastapi.responses import JSONResponse
from serper-dev-api import Serper

app = FastAPI()
serper = Serper(your_api_key)  # Fill in your Serper API Key

async def main():
    async with websockets.connect("ws://localhost:8001") as websocket:
        await websocket.accept()
        print("Connected to MCP Server successfully!")
        
        while True:
            try:
                # Send tool call requests
                tool_name = "search"
                arguments = {"query": "Python tutorials"}
                tool_call = json.dumps({"name": tool_name, "arguments": arguments})
                
                await websocket.send(tool_call)
                response = await websocket.recv()
                print(f"Received response: {response}")
            except Exception as e:
                print(f"Error: {str(e)}")

🚀 Starting Commands

Run the following commands in the terminal to start the Server and Client:

Start the Server

python server.py serve

Start the Client

python client.py main

📚 Documentation

• Model Context Protocol (MCP)
• Serper.dev API
• FastAPI Official Documentation

That's the complete guide on how to build an MCP Server and Client using Python. If you have any questions, please refer to the relevant documentation or seek help in the community!