MCP Server Demonstration

This repository demonstrates how to set up and use MCP (Model Context Protocol) servers within Cursor. MCP allows you to define custom tools that the AI assistant can use to perform actions, such as fetching data from an API or running a local script.

Configuration

The MCP servers are configured in the .cursor/mcp.json file. This file contains a JSON object that defines one or more servers.

{
  "mcpServers": {
    "docker-server_2": {
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "--name",
        "mcp-weather-server",
        "mcp-weather-server:latest"
      ],
      "env": {
        "DEBUG": "mcp:*"
      }
    },
    "time-server_2": {
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "mcp/time"
      ],
      "env": {
        "DEBUG": "mcp:*"
      }
    }
  }
}

Servers

This project defines two MCP servers:

1. docker-server_2: This server runs a Docker container named mcp-weather-server from the mcp-weather-server:latest image. It's used to get weather information.
2. time-server_2: This server runs a Docker container from the mcp/time image. It's used to get the current time in a specific timezone.

Usage

You can interact with these servers by asking the Cursor assistant questions that would trigger the tools defined by these servers.

Getting the Time

User: what time is it in Boston? Assistant Tool Call: mcp_time-server_2_get_current_time({timezone: "America/New_York"})

Getting the Weather

User: what's the temperature in Boston? Assistant: I can get the temperature for you. Do you want it in Celsius or Fahrenheit? User: F Assistant Tool Call: mcp_docker-server_2_get_weather_2({city: "Boston", units: "fahrenheit"})