Your cart (0 items)

Connect your AI assistant to Rad Home

Rad Home has an MCP server, so AI assistants such as Claude and ChatGPT can browse the store for you. They can find categories and products with their prices, read pages such as shipping and returns, and read the blog. Every answer links back to the right page on this website. The server is free and read-only, and needs no account or key.

To have an assistant put an order together for you as well, see AI agent checkout (UCP).

Server address

  • Protocol: Model Context Protocol (MCP) over Streamable HTTP.
  • Sign-in: none. No account or API key is needed.
  • Access: read-only. It can't place orders, change a cart or see accounts.
  • Limits: the same request limits as this website. Too many requests get HTTP 429 with a Retry-After header.

Add it to your AI assistant

Menus change often. If yours look different, look for "Connectors", "MCP servers" or "Tools".

Claude (claude.ai and the Claude apps)

  1. Open Settings > Connectors (Customize > Connectors in some versions).
  2. Choose Add custom connector.
  3. Enter a name, such as Rad Home, and the server address above, then choose Add. Leave the OAuth fields empty.

ChatGPT

  1. Open Settings > Apps & Connectors > Advanced settings and turn on Developer mode. It is only available on some plans.
  2. Choose Create, enter a name and the server address, and set Authentication to No authentication.
  3. In a chat, turn the connector on from the + menu.

Claude Code

claude mcp add --transport http rad-home https://www.radho.me/mcp/store

VS Code (GitHub Copilot)

Add this to .vscode/mcp.json:

{
  "servers": {
    "rad-home": { "type": "http", "url": "https://www.radho.me/mcp/store" }
  }
}

Cursor

Add this to ~/.cursor/mcp.json:

{
  "mcpServers": {
    "rad-home": { "url": "https://www.radho.me/mcp/store" }
  }
}

Other assistants

Any MCP client that connects to remote servers over Streamable HTTP works. Use the server address above with no authentication.

What your assistant can do

ToolWhat it returnsInputs
list_categories The store's categories, each top-level category followed by its subcategories, with their links. page, pageSize (1 to 100, default 50)
list_products_in_category The products in a category with their links, images, short descriptions and the prices on their cards, in US dollars. category (link or slug), page, pageSize (1 to 100, default 20), sort (default, name, price_low_to_high or price_high_to_low; the price orders use the regular list price, as on the website, so sale prices can be out of order)
list_pages The store's information pages that its menus and footer link to. page, pageSize (1 to 100, default 50)
get_page One page's text, such as shipping or returns. url (link or slug)
list_blog_posts The blog's posts, newest first, with their dates and summaries. page (20 posts per page)
get_blog_post One post's text. url (link or slug)

Lists come in pages. Every list result has a paging object with page, pageSize, totalItems, totalPages, hasNextPage and nextPage.

For AI agents

  • Connect to https://www.radho.me/mcp/store with MCP over Streamable HTTP: POST JSON-RPC 2.0 messages with Content-Type: application/json and Accept: application/json, text/event-stream. The server is stateless and issues no session ID.
  • Start with list_categories or list_pages. Lists are paged: request page 1, then paging.nextPage while paging.hasNextPage is true.
  • Every category, product, page and blog post comes with its url on this website. Give the user these urls exactly as returned; never build or guess a url.
  • Prices are in US dollars, as the product cards show them. Options, quantity and stock can change the final price and availability, so send the user to the product's url to check them and to buy. callForPricing means the store gives the price on request.
  • Inputs named url or category take the full link or just its last part (the slug).

Example request:

curl -X POST https://www.radho.me/mcp/store \
  -H "Content-Type: application/json" \
  -H "Accept: application/json, text/event-stream" \
  -d '{"jsonrpc":"2.0","id":1,"method":"tools/call","params":{"name":"list_categories","arguments":{"page":1}}}'