Google's official Analytics MCP server connects an AI client to your GA4 properties (Admin API + Data API), so you can ask questions of your traffic data in plain English — top pages, conversions, channel trends — without building reports. Great for SEO and growth work. Here's the setup.
Step 1 — auth with Application Default Credentials
The server uses Google ADC. Install the gcloud CLI, enable the Analytics APIs in a Google Cloud project, and run:
gcloud auth application-default login
Make sure the authenticated account has access to the GA4 property you want to query.
Step 2 — configure the client
The server is Python-based; run it via uvx/pipx per the repo:
{
"mcpServers": {
"google-analytics": {
"command": "uvx",
"args": ["google-analytics-mcp"]
}
}
}
Requires Python 3.10+ with uv. Restart the client; see MCP config file location.
Verify
Ask: "What were my top 10 landing pages by sessions last 28 days, and which converted best?" Real GA4 numbers back confirm it works.
Common problems
Reauthentication required— ADC expired; rerungcloud auth application-default login.permission denied— the account lacks access to that GA4 property; add it in GA Admin.- Wrong property — specify the GA4 property ID in your question or config.
- No tools in client — see Cursor MCP not working.
Going further
Pair GA4 with Search Console-style workflows, a spreadsheet server (export), and Sequential Thinking for analysis. Browse the data analysis category or curated loadouts.