Google Analytics MCP

Google Analytics 4 holds the answers but hides them behind a confusing interface. This server, published by the Google Analytics team, lets Claude run the reports for you and talk through what changed and why.

It is read-only, so nothing can be broken. Setup is the hard part: you need a Google Cloud project with the Analytics APIs switched on, a one-off sign-in with Google's command-line tool, and Python on your machine. Budget 20 minutes.

It is labelled experimental, which in practice means the setup may change rather than that it is unreliable.

• List your GA4 properties and their settings
• Run standard reports: traffic, pages, channels, conversions
• Build funnel reports
• Check what is happening on the site right now
• Compare periods and explain the change in plain words

Open your terminal in the project folder and run:

claude mcp add analytics-mcp --scope user -e "GOOGLE_APPLICATION_CREDENTIALS=PATH_TO_CREDENTIALS_JSON" -e "GOOGLE_PROJECT_ID=YOUR_PROJECT_ID" -- pipx run analytics-mcp

Or add this to your project's .mcp.json:

{
  "mcpServers": {
    "analytics-mcp": {
      "command": "pipx",
      "args": ["run", "analytics-mcp"],
      "env": {
        "GOOGLE_APPLICATION_CREDENTIALS": "PATH_TO_CREDENTIALS_JSON",
        "GOOGLE_PROJECT_ID": "YOUR_PROJECT_ID"
      }
    }
  }
}

You'll be asked to sign in or paste an API key the first time you use it.

Marketers and site owners who want answers from GA4 without learning Explorations.

• Monthly traffic reports
• Explaining a dip
• Funnel analysis
• Channel comparisons