Understanding the Official MCP Registry

The Official MCP Registry, accessible at registry.modelcontextprotocol.io, serves as a central directory for discovering Model Context Protocol (MCP) servers. It stores metadata about MCP servers, not the server binaries themselves. This means you publish a server.json file that describes your MCP, enabling clients and other directories to find and interact with your service. This approach ensures that the registry remains lightweight and focused on discoverability rather than hosting large datasets or executables.

There are two primary structures for MCP servers that can be listed on the registry: package-based and direct-based. A package-based MCP is essentially a collection of related MCPs that are bundled together for easier distribution and management. A direct-based MCP, on the other hand, refers to a single, standalone MCP server. Understanding which type your MCP is will inform how you structure your server.json file.

For instance, the author recently listed their ThinkReview MCP, a hosted server designed for AI code reviews across platforms like GitHub, GitLab, and Azure DevOps, which processes PR/MR URLs from tools such as Cursor, Claude, and Copilot. This example highlights the practical application of MCPs in streamlining development workflows.

Preparing Your server.json

The core of listing your MCP on the registry is the server.json file. This file acts as an advertisement for your service, providing essential information that clients and other discovery tools will use. The structure of this file depends on whether your MCP is package-based or direct-based.

For Package-Based MCPs

If your MCP is package-based, your server.json file should describe the collection of MCPs. This typically involves listing each individual MCP within the package, along with their respective details. The registry expects a specific format to parse these packages correctly. This includes defining the package name, version, and a list of the MCPs it contains. Each MCP within the package should have its own identifier, version, and a description of its function.

The author's experience with ThinkReview MCP, while a hosted service, illustrates the principle: the metadata needs to clearly articulate what the service does, what platforms it integrates with, and what AI models it supports. For a package-based MCP, this means extending that clarity to each component within the package.

For Direct-Based MCPs

For direct-based MCPs, the server.json file is simpler, describing a single MCP server. Key fields typically include:

  • name: The human-readable name of your MCP server.
  • version: The current version of your MCP server.
  • description: A brief explanation of what your MCP server does.
  • url: The endpoint where clients can access your MCP server.
  • tags: Keywords that help categorize your MCP server (e.g., "code-review", "ai", "github").
  • schema: The schema version your MCP server adheres to, ensuring compatibility.

It is crucial to ensure that the url points to a valid, accessible endpoint for your MCP server. Inaccurate URLs will prevent clients from connecting and utilizing your service.

The author emphasizes that the server.json is not a static document. As your MCP server evolves, you must update its version and any other relevant metadata in the server.json file to reflect these changes. This ensures that discoverability remains accurate and that users are always directed to the most current and functional version of your service.

Publishing Your server.json

Once your server.json file is prepared and validated, the next step is to publish it to the Official MCP Registry. The registry provides a mechanism for submitting these metadata files.

The process generally involves:

  1. Creating a repository: You will typically need to create a Git repository (e.g., on GitHub) to host your server.json file. This allows for version control and easy updates.
  2. Submitting a Pull Request: You then submit a pull request to the Official MCP Registry's repository. This pull request will contain your server.json file.
  3. Review Process: The registry maintainers will review your pull request to ensure that your server.json file adheres to the required schema and contains all necessary information. They will also likely check that the provided URL is valid and points to a functioning MCP server.
  4. Merging: If your submission is approved, the pull request will be merged, and your MCP server will be listed on the registry.

The author's experience suggests that clarity and accuracy in the server.json are paramount for a smooth review process. Any ambiguities or missing information can lead to delays or rejections. It's advisable to consult the registry's documentation for the most up-to-date schema and submission guidelines.

Maintaining Your Listing

Listing your MCP is not a one-time event. To ensure your MCP remains discoverable and relevant, you must maintain your listing. This involves:

  • Updating Versions: Whenever you release a new version of your MCP server, update the version field in your server.json and submit a new pull request.
  • Updating URLs: If the endpoint for your MCP server changes, ensure you update the url field accordingly.
  • Refreshing Descriptions and Tags: If your MCP's functionality expands or its target audience shifts, update the description and tags to reflect these changes.
  • Monitoring for Deprecation: Keep an eye on the registry's announcements for any changes to the MCP protocol or registry requirements.

The author's approach to listing ThinkReview MCP highlights the continuous nature of maintaining a service listing. By keeping the metadata current, you ensure that developers looking for AI code review solutions, for example, can reliably find and integrate with your MCP.

Benefits of Listing

Listing your MCP on the official registry offers significant advantages:

  • Increased Discoverability: Your MCP becomes visible to a wider audience of developers and potential users actively searching for MCP solutions.
  • Enhanced Credibility: Being listed on an official registry lends legitimacy to your MCP server.
  • Integration Opportunities: Other tools and platforms can more easily integrate with your MCP if they can discover it through the registry.
  • Community Growth: A well-maintained registry fosters a stronger MCP ecosystem, benefiting all participants.

The Official MCP Registry acts as a critical piece of infrastructure for the MCP ecosystem. By contributing your MCP's metadata, you not only gain visibility but also help to build out a more robust and interconnected network of AI services. This is akin to how an app store makes mobile applications discoverable and accessible to millions of users; the MCP Registry does the same for AI services built on the Model Context Protocol.