Uh oh!
There was an error while loading.Please reload this page.
- Notifications
You must be signed in to change notification settings - Fork24
Exploring the Model Context Protocol (MCP) through practical guides, clients, and servers I've built while learning about this new protocol.
License
cyanheads/model-context-protocol-resources
Folders and files
| Name | Name | Last commit message | Last commit date | |
|---|---|---|---|---|
Repository files navigation
This repository is a collection of guides, utilities, and server implementations for theModel Context Protocol (MCP) created while learning MCP. It reflects my ongoing exploration and development with this exciting new standard for creating powerful Agent capabilities. Questions and feedback are welcome! 🚀
Disclaimer: The resources in this repository (guides, utilities, server implementations, and the MCP TypeScript Template) were developed independently bycyanheads while exploring the Model Context Protocol. This project is not officially affiliated with the Model Context Protocol organization.Links to official MCP resources are provided below. I appreciate the work of the official MCP team on the specification, SDKs, and documentation!
Introduction |Guides |Utilities |Servers |Getting Started |Official Resources
TheModel Context Protocol (MCP) is an open standard designed to standardize how AI applications (clients/hosts) connect to and interact with external data sources and tools (servers).
Think of it like USB-C for AI: a universal way to plug capabilities into LLM applications.
Key Benefits:
- Consistent Interface: Standardized methods for LLMs to access tools and resources.
- Enhanced Capabilities: Empowers LLMs to interact with databases, APIs, local systems, and more.
- Security & Control: Provides structured access patterns with built-in validation and clear boundaries.
- Extensibility: Easily add new capabilities via servers without modifying core LLM applications.
- Modularity: Develop and maintain specialized functionalities in isolated, reusable server components.
For a more in-depth introduction to MCP, including its design philosophy and technical details, visit the official site:modelcontextprotocol.io.
| Project | Description |
|---|---|
| mcp-ts-template | Provides a beginner-friendly, production-ready template for building MCP servers and clients. Includes essential utilities, examples, and type safety for a solid starting point. |
| Guide | Description |
|---|---|
| MCP Client Development Guide | Learn how to build applications that consume MCP server capabilities. Covers core architecture, lifecycle, tools/resources, security, examples, and advanced topics. (Note: Needs update for latest spec changes) |
| MCP Server Development Guide | Comprehensive guide to building MCP servers. Covers core architecture, protocol fundamentals, server lifecycle, transports (Stdio, Streamable HTTP), building with the TypeScript SDK, defining Tools/Resources/Prompts, advanced features (sampling, roots, streaming, progress, subscriptions, completions), security best practices (updated for Auth Spec2025-03-26), troubleshooting, and example implementations. (Updated May 2025) |
| Cyanhead's MCP 'llms.txt' | A custom llms.txt for faster TypeScript MCP server development using the high-level SDK (McpServer). Tailored for LLM consumption, it covers key concepts, high-level examples, security, and dynamic capabilities, updated for Spec2025-03-26 & TS SDKv1.11.0. |
| Project | Description |
|---|---|
| mcp-reporter | Generates comprehensive capability reports for MCP servers, helping developers understand available functionality across their MCP ecosystem for documentation and integration. |
This repository hosts several example MCP server implementations, showcasing different capabilities:
| Project | Description |
|---|---|
| atlas-mcp-server | ATLAS (Adaptive Task & Logic Automation System), a Neo4j-powered task management system designed for LLM Agents. It uses a three-tier architecture (Projects, Tasks, Knowledge) to manage complex workflows and includes Deep Research scaffolding. |
| clinicaltrialsgov-mcp-server | Enables AI agents to search, retrieve, and analyze clinical study data from ClinicalTrials.gov programmatically via MCP. |
| filesystem-mcp-server | Offers platform-agnostic file system capabilities for AI agents via MCP, enabling file and directory management. Features advanced search/replace and directory traversal. |
| git-mcp-server | Provides an enterprise-ready MCP interface for Git operations, allowing LLMs to initialize, clone, branch, commit, and manage repositories via STDIO & Streamable HTTP. |
| github-mcp-server | Integrates with the GitHub API via MCP, providing a structured interface for LLM agents to manage repositories, issues, pull requests, code, files, and releases. |
| mentor-mcp-server | Offers AI-powered mentorship via MCP using the Deepseek API, providing LLM agents with a 'second opinion'. Can be used for code review, design critique, writing feedback, and brainstorming. |
| ntfy-mcp-server | Integrates with the ntfy.sh push notification service via MCP, enabling LLMs to send highly customizable notifications to external devices. |
| obsidian-mcp-server | Enables LLMs to securely interact with Obsidian vaults via MCP, offering token-aware tools for managing notes. Facilitates seamless knowledge base management with Properties management. |
| perplexity-mcp-server | Unlocks Perplexity's search-augmented AI capabilities for your LLMs via MCP. Provides access to real-time web information with robust error handling and secure validation. |
| pubmed-mcp-server | Connects AI agents to NCBI's PubMed and E-utilities via MCP, enabling search, retrieval, and analysis of biomedical literature. |
| toolkit-mcp-server | Provides essential system utilities and tools for LLM agents via MCP. Features include IP geolocation, network diagnostics, system monitoring, cryptographic operations, and QR code generation. |
| workflows-mcp-server | Empowers AI agents with a powerful, declarative workflow engine to discover, understand, and execute complex, multi-step workflows defined in simple YAML files. Easy as asking your LLM "Create a new workflow to do xyz using the tools you currently have access to" |
- Explore the Guides: Understand MCP concepts and development approaches using theClient andServer guides.
- Select a Server: Choose one relevant to your needs from theMCP Servers section and follow its specific setup instructions in its repository.
- Connect a Client: Use an existing MCP-compatible client (like Claude Desktop, Cline, etc.) or build your own using theClient Development Guide.
- Experiment & Contribute: Try out the tools and consider contributing via issues or pull requests on the respective project repositories.
Key links to official Model Context Protocol documentation, specifications, and community resources (modelcontextprotocol.io):
| Page / Section | Link Path |
|---|---|
| Introduction | /introduction |
| Server Quickstart | /quickstart/server |
| Specification Home | /specification |
| ↳ Architecture | /specification/architecture |
| ↳ Base Protocol | /specification/basic |
| ↳ Server Features | /specification/server |
| ↳ Client Features | /specification/client |
| ↳ Auth Spec (2025-03-26) | /.../authorization |
| Schema Definition (TS) | spec/.../schema.ts |
| Contributing Guide | CONTRIBUTING.md |
| GitHub Organization | github.com/... |
| Specification Discussions | spec/discussions |
| Organization Discussions | orgs/discussions |
| JSON-RPC 2.0 Spec | jsonrpc.org |
SDKs & Tools:
| Language | SDK Repository |
|---|---|
| TypeScript | typescript-sdk |
| Python | python-sdk |
| Kotlin | kotlin-sdk |
| Java | java-sdk |
| C# | csharp-sdk |
| Tool | Repository |
| MCP Inspector (Debugging tool) | inspector |
About
Exploring the Model Context Protocol (MCP) through practical guides, clients, and servers I've built while learning about this new protocol.
Topics
Resources
License
Uh oh!
There was an error while loading.Please reload this page.
Stars
Watchers
Forks
Releases
Sponsor this project
Uh oh!
There was an error while loading.Please reload this page.
Packages0
Uh oh!
There was an error while loading.Please reload this page.