Where to List Your MCP Server in 2026: The Official MCP Registry, Directories and Discovery Files

🔌 MCP · Updated October 2026 · 4 min read · by Jacob Campbell

A public MCP server only helps if people and assistants can find it. There are three places discovery happens: the official MCP Registry, community directories, and your own site. This guide covers each, with what they ask for, and a short checklist of what to get right before you list anywhere, because a listing that points at a broken or insecure server is worse than none.

Key facts

• Sep 2025 — Official MCP Registry launched in preview
• server.json — The metadata file the registry publishes
• Namespaces — Verified via GitHub or your own domain
• 3 places — Registry, directories, your own site

Before you list: a readiness check

1. The server answers initialize, tools/list and tools/call correctly over Streamable HTTP.
2. Every tool has a clear description, an input schema, and annotations; ideally output schemas too.
3. It passes the MCP server security checklist: Origin checks, input validation, rate limits, safe errors.
4. It has a stable public URL that will not change.

The official MCP Registry

The MCP project launched an official registry in preview in September 2025, at registry.modelcontextprotocol.io. It is a catalogue of publicly available servers, designed to be the upstream source that other registries and clients pull from. You describe your server in a server.json file (name, description, version, and either a package for locally run servers or a remote URL for hosted ones) and publish it with the registry’s command-line publisher.

Names are namespaced and verified. A name like io.github.yourname/server is verified by signing in with GitHub; a name under your own domain, such as com.example/server, is verified by DNS or HTTP proof that you control the domain. Domain namespaces are a good fit for a business, because the name matches your website.

Community directories

Many people find servers through directories and marketplaces that list, describe and sometimes test MCP servers, such as Smithery, Glama, PulseMCP and mcp.so. They differ in what they ask for (a GitHub repository, a hosted URL, a manifest) and in whether they test the server. Read each one’s submission guidelines; some pull from the official registry automatically, which is another reason to publish there first.

Your own site

The listing that never goes stale is the one you control:

• An MCP server card under /.well-known/ (we publish ours at /.well-known/mcp.json).
• A line in llms.txt and agents.txt.
• A mention in your A2A agent card if you publish one.
• A human page explaining what the tools do and how to add the server to Claude or ChatGPT.

Details: agent discovery files. Ours are listed on our agent-ready page.

Writing the description

The same rules as llms.txt: state what the server does in one sentence, who it is for, and what it will not do. Name the tools. Say whether it is read-only or can take actions. If it is for a business, include the business name and location so the listing is unambiguous. See llms.txt examples for the tone.

After listing

Watch your logs for new client names and tool calls (AI crawlers in server logs), keep the version in server.json current when tools change, and re-test after every change. WebAgentScan scores MCP quality and agent security as two of its pillars; see what WebAgentScan checks. If you do not have a server yet, the AI Indexing Pack includes one, and adding an MCP server covers building your own.

Product facts: the AI Indexing Pack

Stated plainly, so it can be checked against the product page:

• What it is: the AI Indexing Pack is a set of files you upload to your own website. It is sold by Built2WinWeb as a one-time purchase, with no subscription.
• What it ships: (1) a working MCP server using the Streamable HTTP transport, with typed output schemas and tool annotations; (2) Markdown content negotiation, so agents can request a clean text version of each page; (3) llms.txt, agents.txt and agent cards; (4) the full /.well-known/ discovery tree; (5) OAuth protected-resource metadata; (6) bot access control that verifies crawler identity; (7) a README with a verification checklist.
• Requirements: PHP 7.4 or newer, Apache or LiteSpeed with .htaccess support, FTP access, and a domain you control. Installation takes about 30 minutes.
• What it is not: it is not a WordPress plugin (that is the separate WordPress edition, which needs WordPress 6.0+ and PHP 8.0+), it does not write content for you, it does not guarantee rankings or citations, and it is not a managed service.
• Free routes in: sites built on AIWebBuilding get the core built in, and a 50-day streak on FreeAIBoost unlocks the pack for one domain. Current prices are on the product pages.

Related reading

• Add an MCP server to your website — Tools, transports, annotations and security for a public site.
• MCP server security checklist — 20 checks for a public MCP endpoint.
• Agent discovery files explained — agents.txt, agent cards, server cards and the .well-known tree.
• What WebAgentScan checks — The six scoring pillars, what each one tests, and how to read the result.
• Find AI crawlers in your server logs — Which bots visit, what they fetch and what they get back.
• How this site is agent-ready — Our own live setup: the MCP endpoint, llms-full.txt, Markdown twins and the .well-known tree.
• AI Indexing Pack — The file-drop kit: MCP server, Markdown pages, llms.txt, discovery files and bot access control for any PHP site.

Frequently asked questions

Is there an official MCP server registry?

Yes. The MCP project launched the official MCP Registry in preview in September 2025 at registry.modelcontextprotocol.io.

What is server.json?

The metadata file published to the official registry describing an MCP server: its name, description, version, and package or remote URL.

How does the registry verify who owns a server name?

Names are namespaced. GitHub-based names are verified by GitHub sign-in; domain-based names are verified by DNS or HTTP proof of control of the domain.

Should I list on directories as well?

Yes. Community directories such as Smithery, Glama, PulseMCP and mcp.so are where many people browse, and some pull from the official registry.

What should I check before listing?

Protocol correctness, clear tool descriptions with schemas and annotations, security basics such as Origin checks and rate limits, and a stable URL.