Demo

You can run and interact with a complete MCP server written in Python that exposes Tools, Resources, and Prompts. This server is ideal for demonstrations and learning how MCP works, with simple decorators and type-driven schemas that make it easy to test with MCP clients.

How to use

You connect to the server using any MCP-compatible client. The server presents three kinds of capabilities: Tools you can invoke to perform calculations or analyses, Resources you can fetch to gain context, and Prompts you can trigger to guide a workflow. Use a client to browse available capabilities, test tools with inputs, fetch resources for context, and start prompted workflows.

How to install

Prerequisites: install the MCP runtime and required tooling.

pip install fastmcp

For MCP Inspector (optional):

npm install -g @modelcontextprotocol/inspector

Additional sections

Starting and using the demo involves running either an interactive inspector, a Claude Desktop integration, or a standalone server.

Option 1 interactive demo with MCP Inspector (recommended) runs the inspector against the Python server.

Linux/Mac

cd mcp-demo
./demo.sh
# Choose option 1 for interactive demo

Standalone server and Claude Desktop integration

Run the server directly, or connect via Claude Desktop using a preconfigured MCP server entry.

Standalone start (Windows or other):

python demo_server.py

Claude Desktop configuration example (Windows path shown):

{
  "mcpServers": {
    "demo-server": {
      "command": "python",
      "args": ["D:\\Demo\\mcp-demo\\demo_server.py"]
    }
  }
}

Available tools

calculate_tip

Calculates a tip amount given a bill total and a tip percentage, returning the tip amount, total, and per-person splits.

analyze_text

Analyzes a text input to produce statistics such as word and character counts and estimated reading time.

convert_temperature

Converts a temperature value between Celsius, Fahrenheit, and Kelvin, returning the converted value and the formula used.