Macfax MCP server

macfax-mcp

MCP server for the used-Mac market, by Macfax. Six tools an AI assistant needs to help someone buy or sell a used Mac:

Every price is labeled by basis. The sold slice is the headline sale estimate: verified sales (with sample size and window) or a modeled estimate marked calibrated with its ratio disclosed. Asking bands come from live listings and are seller listing guidance. The API says so in-band, and we ask you to preserve the labels. Free, no account; rate limits are published at macfax.com/developers.

The hosted server (recommended)

A remote, authless, streamable-HTTP MCP server runs at:

https://macfax.com/mcp

• Claude (claude.ai, paid plans): Settings → Connectors → Add custom connector → paste the URL.
• Gemini (Spark, Google AI Ultra): Settings and help → Connected apps → Add a custom app → paste the URL.
• ChatGPT: Settings → Developer mode → add the URL (directory listing pending review).
• Cursor: Install in Cursor
• VS Code: Install in VS Code
• Gemini CLI: gemini extensions install https://github.com/macfax/macfax-mcp
• Anything with an mcp.json:

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

Registry name: com.macfax/macfax on the official MCP registry.

This package (stdio wrapper)

For stdio-only clients, this repo is a thin local wrapper over the same free HTTP API (OpenAPI):

{
  "mcpServers": {
    "macfax": {
      "command": "npx",
      "args": ["-y", "macfax-mcp"]
    }
  }
}

Optional env:

• MACFAXAPIKEY: a free key raises rate limits about 10x. Mint one with curl -X POST https://macfax.com/api/v1/keys (no email needed).

Run it directly:

npx -y macfax-mcp            # speaks MCP over stdio
npx -y macfax-mcp --version  # print the installed version

Notes for integrators

Shortened. The full README is on GitHub.