Skip to main content
search-console-mcp is an open-source implementation of the Model Context Protocol that gives AI agents direct, structured access to Google Search Console (GSC), Bing Webmaster Tools, and Google Analytics 4 (GA4). Unlike simple API wrappers, this project focuses on providing SEO Intelligence Tools. Instead of just asking an agent to “look at my data,” you can give it tools to “find quick wins,” “detect traffic anomalies,” or “submit URLs instantly.”

​
Key Capabilities (v2.0)

  • Fluent Domain Architecture: Streamlined into 7 core domain entry points (sites_list, sitemaps_list, analytics_query, inspection_inspect, indexing_submit, seo_audit, site_health_check).
  • Parallel Multi-Engine Execution (engine: "all"): Query Google, Bing, and GA4 simultaneously with 50%+ lower latency.
  • 100% Backward Compatibility: All ~96 legacy tool names continue working seamlessly via our Fallback Router. Read Backward Compatibility Guide →
  • Multi-Platform Support: Manage sites and behavior data for Google, Bing, and GA4 in one place.
  • Advanced Analytics: Multi-dimensional analysis, rolling averages, period-over-period comparisons, and drop attribution.
  • SEO Insights: Deterministic detection of cannibalization, Striking Distance keywords, and “Low-Hanging Fruit.”
  • Site Health Check: Automated diagnostics across all your properties — performance trends, sitemap status, and anomaly detection in one call.
  • Instant Indexing: Use IndexNow to instantly notify Bing and other engines of content changes.
  • Sitemap Control: List, submit, and delete sitemaps.
  • URL Inspection & PageSpeed: Check indexing status on Google and Bing, and audit Core Web Vitals directly within your SEO workflow.

​
The Problem

Working with SEO data in LLMs usually involves:
  1. Exporting CSVs.
  2. Uploading them to a chat window.
  3. Hoping the model calculates standard deviations or trends correctly.

​
The Solution

With this MCP server, the agent has a “toolbox.” When you ask “Why did my traffic drop?”, the agent doesn’t guess. It calls analytics_anomalies to check for statistical drops, inspection_inspect to check indexing, and analytics_compare({ mode: "drop_attribution" }) to correlate with algorithm updates and GA4 behavior.

​
Supported Clients

This server works with any MCP-compatible client, including: