Everything you need to discover, understand, and implement Model Context Protocol. Curated servers, verified clients, and pre-built skills. All in one place.
We verify based on GitHub activity, community adoption, and active maintenance. No abandoned projects. 600 verified vs competitors' 17k+ unverified.
Our verification process →
Discover not just servers, but the clients to run them, the skills to enhance them, and the knowledge to implement them.
New servers added, unmaintained ones removed. Your source for the living MCP ecosystem, not static lists.
Explore the blog to understand MCP, best practices, and implementation patterns
Complete guide to deploying MCP servers with SSE/WebSocket for cloud scalability, remote teams, and distributed architec...
Read Full Article →Complete guide to deploying MCP servers for local development, stdio-based communication, and CLI workflows.
Read Full Article →Complete guide to enterprise-grade MCP server deployment with SSO, audit logs, SLAs, and production-ready features.
Read Full Article →Discover the most popular and powerful MCP servers, hand-picked based on community adoption and GitHub stars
Quick answers to common questions about My MCP Shelf
We evaluate GitHub activity, star count (real adoption), and maintenance status. If it has active development and community usage, it qualifies. We also check for proper documentation, security practices, and working examples. See our full verification process in the Quality Standards guide.
Quality over quantity. We skip abandoned projects and low-effort forks so you can find what actually works. Every server in our directory has been verified for active maintenance, community adoption, and proper documentation. Other directories include everything ever created—we curate only what's worth using and maintainable.
MCP Servers (600+): Tools/data sources that expose functionality through MCP. MCP Clients (250+): Applications that connect TO servers. Agent Skills (98): Pre-built capabilities that enhance AI agents. Servers provide tools, clients access them, and skills enhance the AI's ability to use them effectively.
Check our MCP Client Selection Guide for detailed recommendations. Key options: Claude Desktop (official, easy), Cursor (for coding), Cline (open source, terminal), Continue (IDE integrated). Consider your use case, OS, and whether you need multiple server connections.
Yes! Click "Submit Your Server" or email us your repository. Requirements: 10+ GitHub stars, active maintenance, proper docs, and working examples. We review submissions weekly.
Have more questions? Check our complete FAQ with 26+ Q&A
View Full FAQ →
