Awesome MCP › Other MCP Servers

surendranb/google-analytics-mcp

⭐ 242 Python repository created 2025-05-24

Google Analytics 4 MCP Server exposes GA4 reporting data to MCP clients such as Claude, ChatGPT, Gemini, Cursor and VS Code. It authenticates with a Google Cloud service account key that has Viewer access on the GA4 property, reads the credentials path and property id from environment variables, and calls the Google Analytics Data API on behalf of the model. Six tools are documented: get_ga4_data runs a query with dimensions, metrics, date ranges and optional filters; search_schema does a keyword search across more than 200 GA4 dimension and metric API names; get_property_schema lists the dimensions and metrics actually available for the configured property; list_metric_categories and list_dimension_categories browse those fields by grouping such as user, session, revenue, event, geography, traffic source and device; and get_troubleshooting_guide returns guidance on IAM permissions, setup and filter syntax. Several behaviours target common model failures. Metric auto-aliasing rewrites legacy or guessed names, for example conversions to keyEvents, so a query does not fail outright. System directives detect schema hallucinations and incorrect filter nesting and steer the model to the troubleshooting tool. Aggregation happens server side for non-time-series questions so the model receives totals rather than raw rows, and a row-count estimate runs before large queries, above roughly 2,500 rows, to avoid flooding the context window. Distribution is broad: a shell installer that auto-configures several clients, a Homebrew tap, an npm package run through npx, a PyPI package run through uvx or python -m ga4_mcp, and a Gemini CLI extension. Anonymous usage telemetry is collected and can be disabled with DISABLE_TELEMETRY or DO_NOT_TRACK. Apache 2.0 licensed.

https://github.com/surendranb/google-analytics-mcp

mcp-servergoogle-analyticsga4analyticsreportingmarketingpythondata-api