What is Boot Mcp

boot-mcp is a starter template designed for building Model Context Protocol (MCP) applications using TypeScript. It provides a structured foundation for developers to create servers that securely expose data and functionality for LLM applications.

Use cases

Use cases for boot-mcp include building command-line tools that interact with LLMs, creating web APIs for LLM applications, and developing applications that require dynamic data loading and execution of code through LLMs.

How to use

To use boot-mcp, install the dependencies using npm, pnpm, or yarn. You can start the example servers by running ‘pnpm start’ for the stdio server or ‘pnpm start:http’ for the HTTP server. The project structure includes directories for server implementations, resources, tools, and prompts.

Key features

Key features of boot-mcp include support for multiple transport methods (stdio and HTTP with SSE), a modular project structure, reusable prompt templates, and the ability to expose data through Resources and functionality through Tools.

Where to use

boot-mcp can be used in various fields where LLM applications are deployed, including AI development, data processing, and any domain requiring secure and standardized server-client interactions.

boot-mcp

A comprehensive starter template for building Model Context Protocol (MCP) applications with TypeScript.

What is MCP?

The Model Context Protocol (MCP) lets you build servers that expose data and functionality to LLM applications in a secure, standardized way. Think of it like a web API, but specifically designed for LLM interactions. MCP servers can:

• Expose data through Resources (think of these sort of like GET endpoints; they are used to load information into the LLM’s context)
• Provide functionality through Tools (sort of like POST endpoints; they are used to execute code or otherwise produce a side effect)
• Define interaction patterns through Prompts (reusable templates for LLM interactions)
• Set boundaries with Roots (defining the operational scope for servers)
• Enable agent behaviors with Sampling (allowing servers to request completions from LLMs)
• Establish connections via Transports (communication mechanisms between clients and servers)

Installation

# npm
npm install

# pnpm
pnpm install

# yarn
yarn install

Quick Start

This template includes examples of MCP servers with different transports:

• Stdio transport (for command-line tools and direct integrations)
• HTTP with SSE transport (for remote servers)

Running the Example Server

# Start the stdio server
pnpm start:basic

# Start the HTTP server
pnpm start:http

Project Structure

boot-mcp/
├── src/
│   ├── index.ts          # Main entry point
│   ├── server/           # MCP server implementations
│   │   ├── basic.ts      # Basic server with common functionality
│   │   └── http.ts       # HTTP with SSE transport server
│   ├── resources/        # Resource implementations
│   │   └── index.ts      # Common resource types and utilities
│   ├── tools/            # Tool implementations
│   │   └── index.ts      # Various tools organized by category
│   ├── prompts/          # Prompt implementations
│   │   └── index.ts      # Common prompt templates and workflows
│   ├── roots/            # Root management
│   │   └── index.ts      # Root definition and utilities
│   ├── transports/       # Transport implementations
│   │   └── index.ts      # Stdio and HTTP/SSE transports
│   └── sampling/         # Sampling utilities (experimental)
│       └── index.ts      # LLM sampling capabilities
├── examples/             # Example usage
│   ├── basic-server.ts   # Basic stdio server example
│   └── http-server.ts    # HTTP server example
└── test/                 # Tests

Features

Resources

• Text and binary resource support
• Dynamic resource templates with parameters
• Standard resource types (files, system info, etc.)
• Resource content helpers

Tools

• System operation tools (execute commands, read files, etc.)
• Data processing tools (JSON parsing, CSV analysis, etc.)
• Utility tools (random generators, string transformations, etc.)
• Calculator and basic tools

Prompts

• Simple greeting and text processing prompts
• Code review prompts with language detection
• Multi-step debugging workflows
• Resource-based prompt templates

Roots

• File system root management
• URI validation and filtering
• Standard root templates

Transports

• Stdio transport for terminal applications
• HTTP/SSE transport for web applications
• Unified configuration interface

Sampling (Experimental)

• Text completion requests
• Conversation-based sampling
• Model preference controls
• Agent workflow examples

License

MIT License © Kirk Lin