PostLaunchKit

What PostLaunchKit MCP server does

The PostLaunchKit MCP server gives AI agents access to the PostLaunchKit directory, which contains human-reviewed free tools and projects, including AI agent tooling. It is hosted by Team Handyapps and is available at https://postlaunchkit.com/mcp.

Agents can search published entries by query, category, tag, and page; retrieve a single entry by slug; and list categories with their entry counts. The service also supports product submissions, website launch checks, and research into sites where an indie product may be listed.

How it works

The PostLaunchKit MCP server uses Streamable HTTP with JSON-RPC requests sent by POST. It is stateless and does not create sessions. Authentication is not required, and the hosted service is free to use.

MCP clients connect directly to the endpoint rather than starting a local process. A compatible configuration uses the server URL under an MCP server name, for example:

{
  "mcpServers": {
    "postlaunchkit": {
      "url": "https://postlaunchkit.com/mcp"
    }
  }
}

The endpoint can also be checked with a JSON-RPC tools/list request using curl.

Setup and configuration

There is no package to install for the PostLaunchKit MCP server. Add the hosted URL to an MCP client that supports remote servers and Streamable HTTP. The repository documents the service rather than containing a separately run implementation.

No environment variables, API keys, or authentication settings are specified. The project also provides an agent guide, an OpenAPI description for its REST API, an llms.txt file, and the public directory as additional ways to access information.

Tools and capabilities

• search_projects searches published directory entries with optional query, category, tag, and page filters.
• get_project returns one published entry when given its required slug.
• list_categories returns categories and the number of entries in each.
• submit_project sends a product suggestion to a pending review queue. Required fields are name, URL, one-line description, reason to try it, and category; tags, contact information, and agent name are optional.
• run_launch_check evaluates a public website and reports the number of checks passed, the pass or fail result for each check, suggested fixes for failures, and a report link.
• list_launch_directories returns directory and listing-site data from the open launch-directories dataset. Optional filters cover free listings, dofollow links, account requirements, AI-agent friendliness, and text search.

Limitations and notes

Submitting a product does not publish it automatically. A person reviews submissions before publication, and duplicate domains are rejected. Submission volume is limited by IP address and day.

Launch checks accept public websites only and are limited to 20 checks per IP per day. In the launch-directory dataset, blank fields indicate that a value has not been verified. That dataset is available under CC BY 4.0.

The project is licensed under MIT. The service is intended for directory lookup, submission, launch-check results, and launch-site research; the provided material does not describe tools for modifying published entries or managing review decisions.