Home Getting Started Connect an AI Assistant (MCP)

Connect an AI Assistant (MCP)

Last updated on Sep 04, 2026

Ranksy runs an MCP server, so you can connect an AI assistant — Claude Desktop, Claude Code, or any Model Context Protocol client — and ask it questions about your live Ranksy data in plain language.

What You Can Ask

Once connected, the assistant reads from your live Ranksy data. You can ask things like:

  • "What are my top-moving keywords this week?"
  • "How did my rankings change after my last listing update?"
  • "Which competitors gained reviews in the last 30 days?"
  • "What's my MRR and churn trend this quarter?"
  • "Which keywords am I being cannibalized on?"

You ask in your own words; the assistant fetches the numbers and answers. It's the same data you see in the dashboard, reachable from wherever your AI assistant lives.

What You'll Need

  • A Ranksy account (you'll authorize the connection with it).
  • An MCP client — Claude Desktop, Claude Code, or another Model Context Protocol host.
  • Node.js installed, for the small mcp-remote bridge that connects your client to Ranksy.

Get Your MCP Server URL

  1. Go to Settings → AI Assistants.
  2. Copy your MCP Server URL. It looks like https://ranksyapp.com/mcp.

That page also shows your usage, connected assistants, and a ready-to-copy Claude Desktop config.

Connect Claude Desktop

  1. Open (or create) your Claude Desktop config file:
    • macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
    • Windows: %APPDATA%\Claude\claude_desktop_config.json
  2. Add Ranksy to it:
{
  "mcpServers": {
    "ranksy": {
      "command": "npx",
      "args": ["-y", "mcp-remote", "https://ranksyapp.com/mcp"]
    }
  }
}
  1. Save the file and restart Claude Desktop.
  2. When prompted, authorize the connection in your browser. That's it — no key to paste.

💡 The Settings → AI Assistants page has this exact config pre-filled with your URL. Copy it from there instead of typing it by hand.

Quick Test (Terminal)

To check the connection without an app, run this in your terminal:

npx -y mcp-remote https://ranksyapp.com/mcp

It starts an MCP client and opens your browser to authenticate. Node.js is required.

Managing Connections

Every assistant you connect appears under Connected Assistants on the AI Assistants page, with when it connected and when its access expires.

  • Connecting uses a browser authorization (OAuth) — there's no long-lived key sitting in a config file, just the connection itself.
  • To disconnect an assistant, hit the trash icon next to it and confirm. It loses access to your Ranksy data immediately.

Usage and Limits

Tool calls — each time the assistant actually fetches data — are metered per billing cycle, and your plan sets the limit. Connecting and listing the available tools is free. You can watch your usage, a per-tool breakdown, and recent activity right on the AI Assistants page.

MCP vs the Developer API

Both let you get at your Ranksy data, but they're for different jobs:

  • AI Assistant (MCP) — you, exploring your data conversationally. Browser authorization, no code.
  • Developer API — your app or scripts, reading and writing data programmatically. API keys, REST.

Use the assistant to ask questions; use the API to build on top of Ranksy.

FAQ

Do I need to paste an API key? No. MCP authorizes in the browser. API keys are a separate thing, for the Developer API.

Which assistants work? Any Model Context Protocol client. Claude Desktop and Claude Code are the common ones; the mcp-remote bridge connects the rest.

Is my data safe? An assistant only gets access after you authorize it, and only reads the data your account can already see. Revoke it anytime from Connected Assistants.

Why does it need Node.js? The mcp-remote bridge runs on Node. It's what carries the connection between your assistant and Ranksy.

Next Steps