abhay-2108/Remote-MCP-Server-Testing

MCP Demo Project

This project demonstrates setting up and running an MCP (Model Context Protocol) server using FastMCP and uv.

Prerequisites

• Python 3.11+
• uv CLI tool
• fastmcp Python package
• GitHub account (for repository)
• FastMCP Cloud account (free tier)

Setup Instructions

1. Initialize the project using uv

   pip install uv
   uv init .
   

2. Add FastMCP to the project

   uv add fastmcp
   

3. Create your MCP server file (e.g., main.py)

   from fastmcp import FastMCP
   
   mcp = FastMCP("DemoServer")
   
   @mcp.tool
   def add(a: int, b: int) -> int:
       return a + b
   
   if __name__ == "__main__":
       mcp.run() 
       * To run MCP server locally. Default is to run locally
       mcp.run(transport="http",host="0.0.0.0",port=8000)
       * To run MCP server on FastMCP Cloud.
   

4. Start the MCP server locally

   uv run fastmcp run main.py
   

   Start the Remote MCP server

   fastmcp run main.py --transport http --host 0.0.0.0 --port
   
                              or 
   
   uv run main.py
   

5. Test server using MCP Inspector

   uv run fastmcp dev main.py
   

   • Allows you to test and inspect MCP tools locally (similar to Postman).

Deploy MCP Server on FastMCP Cloud as a Remote Server

1. Create a GitHub repository for your project and push your code.

2. Create a FastMCP Cloud account (free tier available at FastMCP Cloud).

3. Install project dependencies on cloud

   uv add .
   uv add fastmcp
   

4. Deploy your MCP server

   uv run fastmcp deploy main.py --cloud
   

5. Test deployed server using MCP Inspector or via your cloud URL.

6. To integrate MCP server with Claude Desktop powershell uv run fastmcp install claude-desktop main.py

• Integrating MCP server with Claude Desktop is straightforward.

Using Proxy Server / Remote MCP Server Locally

You can connect to a remote MCP server using a proxy to access MCP tools locally. This is useful for testing cloud-deployed MCP servers or running your server from a remote environment.

Steps

1. Ensure your remote MCP server is deployed and accessible (for example, on FastMCP Cloud).
2. Use the as_proxy method to create a proxy client in your local environment.

Example: Using as_proxy Method

from fastmcp import FastMCP

# Connect to the remote MCP server using proxy
mcp = FastMCP.as_proxy("https://your-remote-mcp-server.fastmcp.ai")

# Call remote MCP tools as if they were local
result = mcp.add_expense(item="Lunch", amount=12.5, category="Food")
print(result)

summary = mcp.expense_summary()
print(summary)

This setup allows you to interact with the remote MCP server seamlessly, as

Project Structure

MCP Demo Project/
├─ main.py          # MCP server code
├─ uv.yaml          # Project config (generated by uv)
├─ README.md        # Project documentation
└─ .venv/           # Project virtual environment (managed by uv)

Notes

• Use uv run fastmcp dev main.py for development and testing.
• Use uv run fastmcp run main.py for running the production MCP server locally.
• Deploy on FastMCP Cloud using uv run fastmcp deploy main.py --cloud.
• All MCP tools are defined in main.py under @mcp.tool decorators.