A curated Chinese-language list of MCP (Model Context Protocol) resources, covering MCP basics, clients, and categorized servers for Chinese-speaking developers.
modelcontextprotocol/registry
A community-run registry service where MCP clients can discover MCP servers, acting like an app store for them.
- Stars
- 7,394
- Forks
- 1,033
- Open issues
- 193
- Last commit
- 18 hours ago
What is it?
The MCP Registry is a service that gives MCP clients a list of available MCP servers, functioning like an app store for the Model Context Protocol ecosystem. Server authors publish their servers to the registry so that clients can find them in one place.
Publishing is done with the mcp-publisher CLI from the server's own project rather than through a pull request to this repository. Publishers authenticate through GitHub OAuth or GitHub OIDC in GitHub Actions, or prove domain ownership via DNS or HTTP challenges, and the registry checks namespace ownership when a server is published.
The project is written in Go and runs as an HTTP API backed by PostgreSQL. A local development environment can be started with a single make command using Docker Compose, and pre-built container images are published to GitHub Container Registry. The API is in a v0.1 freeze, with no breaking changes planned for the near term.
Key features
- 01Lists MCP servers that MCP clients can query and discover
- 02Publishing via the mcp-publisher CLI from each server's own project
- 03Namespace ownership checks through GitHub login, GitHub OIDC, DNS, or HTTP verification
- 04Go HTTP API backed by PostgreSQL
- 05Local development environment via Docker Compose and ko-built images
Use cases
- Publishing an MCP server so that clients can discover it
- Building an MCP client that needs to browse available servers
- Integrating registry lookups into tooling or client applications
- Running a private or self-hosted registry for development and testing
Quick start
- 1
Install Docker, Go, ko and golangci-lint (see README prerequisites)
- 2
Start the full development environment with PostgreSQL
make dev-compose - 3
Build the publisher CLI (optional, for publishing a server)
make publisher
Things to know
- Publishing a server is not done through pull requests to this repository; use the mcp-publisher CLI instead.
- The API is in preview and the database uses ephemeral storage that resets on container restart in development.
- The license is listed as NOASSERTION, so the license terms are unclear from the metadata.
In collections
Topics
FAQ
- What is registry?
- A community-run registry service where MCP clients can discover MCP servers, acting like an app store for them. The MCP Registry is a service that gives MCP clients a list of available MCP servers, functioning like an app store for the Model Context Protocol ecosystem. Server authors publish their servers to the registry so that clients can find them in one place.
- Is registry free to use? What license does it have?
- registry is open source under the Other license. Check the license terms before using it in commercial products.
- Is registry still maintained?
- Actively maintained. The last commit to registry was on Oct 21, 2026.
- What can I use registry for?
- Publishing an MCP server so that clients can discover it Building an MCP client that needs to browse available servers Integrating registry lookups into tooling or client applications Running a private or self-hosted registry for development and testing
- How do I install registry?
- Install Docker, Go, ko and golangci-lint (see README prerequisites) Start the full development environment with PostgreSQL: make dev-compose Build the publisher CLI (optional, for publishing a server): make publisher
- Where can I get registry?
- The source code and installation instructions are on GitHub: https://github.com/modelcontextprotocol/registry
Summary generated by AI from the project's README. GitHub stats refresh daily.
Related projects
A framework that gives LLM-based AI agents standardized access to 1000+ scientific tools, datasets and APIs for research workflows.
A Python framework for building MCP servers, clients, and interactive apps that connect LLMs to tools and data.
Open-source platform combining AI agents and workflow automation, self-hostable or embeddable in SaaS products, with 250+ connectors.
A minimal, token-efficient MCP server that lets AI clients query and explore PostgreSQL, MySQL, SQL Server, Oracle, MariaDB and SQLite databases.
Lightweight cross-platform database client for 100+ databases, with desktop, Docker, CLI, a built-in AI SQL assistant and an MCP server.