List your SaaS in the Official MCP Registry

Why this matters

A registry listing puts your MCP server where AI agents and clients look up tools.

Before you start

• A publicly reachable remote MCP server endpoint
• Homebrew installed on macOS or Linux
• An account for the Official MCP Registry (GitHub authentication is typical)

How to apply it

1. Confirm the remote MCP server responds

   Send an MCP initialize request to your public endpoint and check that it answers with a JSON-RPC result containing serverInfo.

   Check: The curl request returns HTTP 200 and a JSON body with a result.serverInfo object.

2. Install the publisher CLI

   Install mcp-publisher with Homebrew as described in the tweet.

   Check: `mcp-publisher --help` exits with status 0 and prints the available subcommands.

3. Authenticate the CLI General best practice

   General best practice: sign the CLI in to the registry account (typically GitHub-based) before any publish attempt, so the submission is attributed to your organisation.

   Check: The CLI reports a signed-in account or writes a credentials file, and no authentication error is returned.

4. Publish the server entry General best practice

   General best practice: run the publish flow from the repository that holds the server manifest, with the name, short description, repository URL and the public remote endpoint filled in.

   Check: The CLI prints a success message naming the published server, with no validation errors.

5. Check the public listing General best practice

   Open the registry page for your server and confirm the description, repository link and endpoint are correct and that search on the registry finds it by name.

   Check: The registry page for your server returns HTTP 200, shows your remote endpoint, and the site name is returned in registry search results.

<goal>
audit link cluster payload image redirect heading route serp audit locale report cluster template.
</goal>

<context>
cluster cluster section crawler content heading template report effort canonical schema metadata alt page.
</context>

<discover_first>
redirect locale canonical heading route crawl snippet canonical template serp signal pipeline schema index.
</discover_first>

<constraints>
crawl render priority index signal section crawler canonical anchor schema crawl layout intent alt.
</constraints>

<plan>
crawl pipeline payload index heading intent intent pipeline heading report priority layout layout effort.
</plan>

<verify>
signal verify report section query alt render serp anchor render layout markup markup snippet.
</verify>

<report>
priority content heading crawl image page heading anchor index cluster cluster content template pipeline.
</report>

Source: @tarasshyn on 2026-10-18 · original post · report an error