MCP Guide for AI Agents

What Is MCP?

The Model Context Protocol (MCP) is an open standard that lets AI assistants connect to external services. When you connect an AI tool like Claude, Cursor, or Windsurf to Mane Match via MCP, the AI can search our horse listings, look up breed statistics, and pull seller profiles on your behalf — all through a structured, machine-friendly interface.

No API key or account is required. The MCP server is read-only and publicly accessible.

Connecting

The Mane Match MCP server is available at:

POST https://mane-match.com/mcp

It uses the Streamable HTTP transport (JSON-RPC over HTTP). Most MCP-compatible tools handle the connection details automatically. To connect, add the following to your MCP client configuration:

{
  "mcpServers": {
    "mane-match": {
      "url": "https://mane-match.com/mcp"
    }
  }
}

For local development, replace the URL with http://localhost:3030/mcp.

Available Tools

These tools let an AI assistant query Mane Match on your behalf. All tools are read-only.

search_listings

Search active horse listings with filters. Supports partial matching on breed, discipline, and location. Pagination is supported via limit and offset.

Parameters: breed, discipline, minPrice, maxPrice, location, gender, minAge, maxAge, limit (1–50, default 20), offset.

get_listing

Retrieve full details for a single listing by its numeric ID. Returns all listing fields including photos, videos, pedigree information, and seller details.

Parameters: id (required).

browse_breeds

List every horse breed currently represented on the marketplace, along with the number of active listings for each. Sorted by count, most popular first.

Parameters: none.

browse_disciplines

List every riding discipline on the marketplace with active listing counts. Sorted by count.

Parameters: none.

get_seller_profile

Look up a seller by email address or vanity slug. Returns the seller's display name, verification status, business information, bio, and a list of their active listings.

Parameters: identifier (email or slug, required).

Available Resources

Resources provide structured data that an AI can read without calling a tool.

• mane-match://listings/catalog — a JSON summary of up to 200 active listings with basic fields (name, price, breed, discipline, location).
• mane-match://breeds — all breeds with listing counts as JSON.
• mane-match://info — general marketplace info including live counts of active listings and distinct breeds, plus contact details.

REST API

If your tool does not support MCP, the same listing data is available as plain JSON. No authentication is required.

• All listings: GET /api/listings
• Single listing: GET /api/listings/{id}
• Listing comments: GET /api/listings/{id}/comments

Agent Policy

Bots, crawlers, AI agents, and automated tools are welcome on Mane Match. We ask that you follow these guidelines:

• Be respectful of resources. Do not send more than 60 requests per minute to any single endpoint. The MCP server and REST API are not rate-limited for normal use, but excessive traffic may be throttled.
• Use the MCP server or JSON API when possible. Scraping HTML pages is unnecessary since all listing data is available in structured form through the MCP server, the REST API, and the sitemap.
• Do not create accounts or post content automatically. Account registration, listing creation, messaging, and other write operations are intended for human users. Automated access to these features is not permitted without prior written authorization.
• Respect robots.txt directives. Authenticated pages (dashboard, admin, settings, messages) are disallowed for crawlers. Do not index or cache content from those paths.
• Identify yourself. If you are building a bot or agent that accesses Mane Match regularly, include a descriptive User-Agent string so we can identify your traffic and reach out if there are issues.
• Cache responsibly. Listing data changes frequently. We recommend caching listing results for no longer than 15 minutes and breed or discipline statistics for up to 1 hour.
• Do not use listing data to contact sellers outside of Mane Match. Seller contact information is provided for buyers who are genuinely interested in a horse. Harvesting emails or phone numbers for marketing, spam, or any purpose unrelated to a horse transaction is prohibited.
• Attribution is appreciated. If you surface Mane Match listing data in your application, a link back to the listing page or to mane-match.com is appreciated but not required.

Other Machine-Readable Files

• llms.txt — a plain-text description of the site, its pages, and API endpoints, written for large language models.
• sitemap.xml — a standard XML sitemap with all indexable URLs and last-modified dates.
• robots.txt — crawler directives.

Questions

If you are building an integration with Mane Match and have questions, reach out to support@mane-match.com.