Gizmoanalytics MCP

What sudomichael/gizmoanalytics-mcp MCP server does

The sudomichael/gizmoanalytics-mcp MCP server exposes Gizmo Analytics operations to AI coding agents through MCP. Gizmo Analytics is a cookieless web analytics service, so the server is intended for adding tracking to websites and querying the resulting data.

Its 28-tool surface covers two main areas. Installation and administration tools can create a site, produce an installation snippet, configure event tracking, add site tags, and work with goals or funnels. Analysis tools can summarize multiple sites, show workspace information, inspect site traffic, rank sites by growth, identify inactive sites, detect anomalies, explain traffic changes, and query dimensions such as pages, referrers, countries, devices, browsers, channels, cities, and AI sources.

How it works

The sudomichael/gizmoanalytics-mcp MCP server is hosted by Gizmo Analytics at https://gizmoanalytics.io/mcp; the repository documents the public MCP listing and does not provide a local runtime. An MCP-compatible client connects to that URL and discovers the available tools.

OAuth-aware clients can discover the authorization service through RFC 9728 metadata, open a browser consent flow, and store the resulting token. Clients without OAuth-aware discovery can send an Authorization: Bearer header containing an admin key. Tool requests then operate against the authenticated Gizmo Analytics workspace.

The installation workflow can combine site creation, a framework-specific tracking snippet, a suggested file location, and next steps. A verification operation checks whether the script is loading by fetching the site homepage. Query operations support custom date ranges and stacked filters, including combinations such as country and device filters.

Setup and configuration

To use the sudomichael/gizmoanalytics-mcp MCP server, add the hosted URL as an MCP server in the client’s MCP settings. Cursor, Claude Desktop, and Windsurf are named as clients that can use automatic discovery. No local package installation or local server process is required.

For manual configuration, place the hosted URL in the client’s MCP server configuration and provide a bearer token containing an admin key. The README directs users to the Gizmo Analytics dashboard’s keys page to obtain that key. OAuth is the preferred setup when the client supports MCP authorization discovery.

Tools and capabilities

Representative tools include:

• create_site for adding a site.
• get_install_snippet for retrieving tracking code.
• summarize_all_sites for a cross-site summary.
• get_workspace_overview for workspace-level information.
• get_site_traffic for site traffic data.
• setup_analytics_for_site for a combined installation workflow.
• verify_install for checking whether tracking loads.

The broader inventory also includes event and badge snippets, site tags, goals, funnels, time-series queries, event queries, property breakdowns, top-page and referrer queries, and workflow tools for growth, anomalies, and traffic changes. Framework-aware installation is described for Next.js, Remix, SvelteKit, Nuxt, Astro, Vite, React, and plain HTML.

Limitations and notes

This repository is not a local MCP implementation that can be started from the checkout. It identifies and documents the hosted service, so use the hosted endpoint rather than looking for a local launch command. Authentication is required through OAuth or an admin bearer key. The provided material does not specify service pricing, data retention, rate limits, or detailed query schemas.