Sponsored by Deepsite.site

MCP Template

Created By
adamwulf9 months ago
A barebones MCP server implementation in Swift using loopwork-ai's mcp-swift-sdk.
Content

MCP Template

A template repository providing a barebones foundation for building Model Control Protocol (MCP) servers for macOS applications and command line tools.

Overview

MCP Template serves as a starting point for developers looking to implement MCP servers in their projects. This template demonstrates how to use the mcp-swift-sdk in a minimal way, making it easier to understand the basics of MCP integration. It includes both a library template for integration into other projects and a simple command-line example to illustrate basic usage.

Purpose

This repository is intended to be:

  • A reference implementation showing how to use mcp-swift-sdk
  • A template that can be forked or cloned as a foundation for your own MCP server implementations
  • A barebones example that demonstrates core MCP concepts with minimal code

Features

Current and planned features include:

  • Basic Swift package structure
  • Command line "hello world" example tool
  • Command line stdio for direct MCP interaction via the run command
  • App Store safe command line stdio → standalone Mac app communication
  • SSE server in a Package → example command line app for SSE-based MCP

Installation

Swift Package Manager

Add the following to your Package.swift file:

dependencies: [
    .package(url: "https://github.com/your-username/mcp-template.git", branch: "main"),
]

Then add the dependency to your target:

.target(
    name: "YourTarget",
    dependencies: [
        .product(name: "EasyMCP", package: "mcp-template")
    ]
),

Usage

Basic Example

import EasyMCP

// Create an instance of EasyMCP
let mcp = EasyMCP()

// Register a tool
try await mcp.register(tool: Tool(
    name: "helloPerson",
    description: "Returns a friendly greeting message",
    inputSchema: [
        "type": "object",
        "properties": [
            "name": [
                "type": "string",
                "description": "Name of the person to say hello to",
            ]
        ],
        "required": ["name"]
    ]
)) { input in
    // It's an async closure, so you can await whatever you need to for long running tasks
    await someOtherAsyncStuffIfYouWant()
    // Return your result and flag if it is/not an error
    return Result(content: [.text(hello(input["name"]?.stringValue ?? "world"))], isError: false)
}

// Start the MCP server for full MCP interaction
try await mcp.start()
try await mcp.waitUntilComplete()

Command Line Example

The package includes a command line executable called mcpexample that demonstrates basic usage:

# Run the basic hello example
mcpexample hello

# Start the MCP server to handle MCP protocol communications
mcpexample run

Project Structure

Targets

  1. EasyMCP (Library)

    • Minimal template implementation of an MCP server
    • Demonstrates basic integration with the MCP protocol
    • Shows how to leverage the official mcp-swift-sdk
    • Includes a simple tool example (helloworld)
  2. mcpexample (Executable)

    • Simple command-line example using the EasyMCP library
    • Includes both a hello command and a run command
    • The run command starts a full MCP server using stdio transport
    • Uses ArgumentParser for CLI argument handling
  3. EasyMCPTests (Test Target)

    • Template tests for the EasyMCP library functionality
    • Includes a basic test for the hello function

Dependencies

  • swift-argument-parser (1.3.0+) - Used for CLI argument handling
  • mcp-swift-sdk (branch: "feature/wait-for-complete") - Custom fork of the MCP implementation that adds the ability to wait for the mcp server to finish

Development

To use this template for your own MCP server:

  1. Clone or fork the repository
  2. Build the package to verify everything works:
    swift build
    
  3. Run the tests:
    swift test
    
  4. Modify the EasyMCP implementation to add your custom functionality
  5. Extend the command line example or create your own Mac application

It may also be helpful to use the MCP Inspector to diagnose and debug a custom MCP server.

$ npx @modelcontextprotocol/inspector

Testing your MCP command line executable

To test and debug your MCP server using the MCP Inspector:

  1. Build your command line executable in Xcode
  2. Locate the executable by going to Xcode → Product → Show Build Folder in Finder
  3. Copy the absolute path of the executable from that directory
  4. Use the MCP Inspector to test your server
  5. Open Terminal and run:
    npx @modelcontextprotocol/inspector <absolute_path_to_your_executable> run
    
  6. Open your browser to the port shown in the output:
    🔍 MCP Inspector is up and running at http://localhost:5173 🚀
    
  7. Press the Connect button in the MCP Inspector interface
  8. Open Activity Monitor and search for your executable name
  9. Verify only the inspector and a single instance of your tool is running
  10. In Xcode → Debug → Attach to Process → Find your executable name at the top and attach
  11. In Terminal, run tail -n 20 -f ~/Library/Logs/Claude/mcp*.log
  12. Now you can interact with your server through the inspector while hitting breakpoints in Xcode!

This setup provides a powerful debugging environment where you can:

  • Test your MCP server's functionality through the Inspector's UI
  • Set breakpoints in your code to trace execution
  • Inspect variables and state during operation
  • Debug resource, prompt, and tool implementations in real-time

For more debugging tips, visit MCP Debugging at Anthropic's modelcontextprotocol.io site.

License

MIT licensed.

Acknowledgments

Recommend Servers
TraeBuild with Free GPT-4.1 & Claude 3.7. Fully MCP-Ready.
Context7Context7 MCP Server -- Up-to-date code documentation for LLMs and AI code editors
Playwright McpPlaywright MCP server
Amap Maps高德地图官方 MCP Server
WindsurfThe new purpose-built IDE to harness magic
TimeA Model Context Protocol server that provides time and timezone conversion capabilities. This server enables LLMs to get current time information and perform timezone conversions using IANA timezone names, with automatic system timezone detection.
AiimagemultistyleA Model Context Protocol (MCP) server for image generation and manipulation using fal.ai's Stable Diffusion model.
MiniMax MCPOfficial MiniMax Model Context Protocol (MCP) server that enables interaction with powerful Text to Speech, image generation and video generation APIs.
Serper MCP ServerA Serper MCP Server
Baidu Map百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Visual Studio Code - Open Source ("Code - OSS")Visual Studio Code
BlenderBlenderMCP connects Blender to Claude AI through the Model Context Protocol (MCP), allowing Claude to directly interact with and control Blender. This integration enables prompt assisted 3D modeling, scene creation, and manipulation.
Zhipu Web SearchZhipu Web Search MCP Server is a search engine specifically designed for large models. It integrates four search engines, allowing users to flexibly compare and switch between them. Building upon the web crawling and ranking capabilities of traditional search engines, it enhances intent recognition capabilities, returning results more suitable for large model processing (such as webpage titles, URLs, summaries, site names, site icons, etc.). This helps AI applications achieve "dynamic knowledge acquisition" and "precise scenario adaptation" capabilities.
Tavily Mcp
EdgeOne Pages MCPAn MCP service designed for deploying HTML content to EdgeOne Pages and obtaining an accessible public URL.
ChatWiseThe second fastest AI chatbot™
MCP AdvisorMCP Advisor & Installation - Use the right MCP server for your needs
Howtocook Mcp基于Anduin2017 / HowToCook (程序员在家做饭指南)的mcp server,帮你推荐菜谱、规划膳食,解决“今天吃什么“的世纪难题; Based on Anduin2017/HowToCook (Programmer's Guide to Cooking at Home), MCP Server helps you recommend recipes, plan meals, and solve the century old problem of "what to eat today"
CursorThe AI Code Editor
Jina AI MCP ToolsA Model Context Protocol (MCP) server that integrates with Jina AI Search Foundation APIs.
DeepChatYour AI Partner on Desktop