Google Analytics 4

Yet another Google Analytics MCP Server that connects your AI assistants with Google Analytics data.

It exposes Google Analytics Admin and Data APIs as callable tools. This allows developers to query account structures, run reports, and retrieve metadata without writing boilerplate API code.

Features

• 📊 Retrieve account, property, and Google Ads link details.
• 📈 Run standard and realtime reports using GA Data API.
• 🔍 Get available dimensions, metrics, and filter hints for accurate queries.
• 🛠️ Auto-hinted parameters reduce errors in report configurations.
• 🔑 Built on Application Default Credentials for secure, local auth.

How to Use It

1. Install pipx via pip:

python -m pip install pipx

2. In your Google Cloud project, enable:

• Google Analytics Admin API
• Google Analytics Data API

3. Authenticate using Application Default Credentials (ADC):

gcloud auth application-default login \
 --scopes https://www.googleapis.com/auth/analytics.readonly,https://www.googleapis.com/auth/cloud-platform \
 --client-id-file=YOUR_CLIENT_JSON_FILE

Replace YOUR_CLIENT_JSON_FILE with your OAuth client JSON. The user must have read access to the GA properties.

Alternatively, use service account impersonation:

gcloud auth application-default login \
 --impersonate-service-account=SERVICE_ACCOUNT_EMAIL \
 --scopes=https://www.googleapis.com/auth/analytics.readonly,https://www.googleapis.com/auth/cloud-platform

4. Configure your MCP clients like Claude, Gemini CLI, or Cursor:

{
 "mcpServers": {
   "analytics-mcp": {
     "command": "pipx",
     "args": [
     "run",
     "--spec",
     "git+https://github.com/googleanalytics/google-analytics-mcp.git",
     "google-analytics-mcp"
   ]
 }
}
}

Optional: Pin credentials by adding the GOOGLE_APPLICATION_CREDENTIALS env variable:

"env": {
 "GOOGLE_APPLICATION_CREDENTIALS": "/path/to/your/adc.json"
}

5. Open your AI assistant and type /mcp. You should see analytics-mcp listed. Try prompts like:

• what can the analytics-mcp server do?
• run a report for users and sessions last 7 days
• list custom metrics in my property with 'web' in the name

Available Tools

• get_account_summaries: Lists all GA accounts and properties accessible to the user.
• get_property_details: Fetch property settings like timezone, currency, and industry category.
• list_google_ads_links: Returns linked Google Ads accounts for a given property.
• run_report: Executes a report with dimensions, metrics, date ranges, and filters.
• get_dimensions, get_metrics: Retrieve all available dimensions/metrics for a property (including custom ones).
• get_standard_dimensions, get_standard_metrics: List built-in GA dimensions and metrics.
• run_report_date_ranges_hints: Suggests valid date range formats for the run_report tool.
• run_report_metric_filter_hints, run_report_dimension_filter_hints: Help construct correct filter syntax.
• run_realtime_report: Pulls live user count, active events, and current platforms.
• get_realtime_dimensions, get_realtime_metrics: Lists what you can query in real-time.

FAQs

Q: Can I use this without a Google Cloud project?
A: No. You need a Google Cloud project with the Analytics Admin and Data APIs enabled. The server relies on Google’s auth system and API endpoints, which require project-level configuration.

Q: Does this support GA4 only, or Universal Analytics too?
A: GA4 only. The server uses the GA Data API v1, which doesn’t support Universal Analytics. UA properties won’t appear in results.

Q: Why am I getting permission errors even with ADC set up?
A: ADC must include the https://www.googleapis.com/auth/analytics.readonly scope. Also, the authenticated user needs Read permission on the GA property. Double-check IAM roles in Google Cloud and GA account access.

Q: Can I run this server headless in a CI/CD pipeline?
A: Yes, but use a service account key file. Set GOOGLE_APPLICATION_CREDENTIALS to the key path and ensure the service account has the required GA read permissions. Avoid user-based ADC in automated environments.

Q: Is the server stateful or does it make live API calls each time?
A: It makes live API calls. No caching or local storage. Each tool invocation hits Google’s backend, so results are current but subject to rate limits.