Setting Up Google Search Console MCP: Give Claude Access to Your Search Performance Data

Stuart Clark avatar
Setting Up Google Search Console MCP: Give Claude Access to Your Search Performance Data

AI-Powered SEO & CRO  ·  stuseo.uk

PART 3 OF 9

Setting Up Google Search Console MCP

Give Claude access to your ranking positions, CTR, and indexing data

AI-POWERED SEO & CRO WITH CLAUDE — 9-PART SERIES

You are reading Part 3 of 9. Start from the beginning if you haven’t read the introduction yet.

Search Console Is Your Most Underused SEO Asset

Google Search Console contains data that’s genuinely unique — you can’t get it anywhere else. Your actual click-through rates by query. The impressions your site gets for keywords you didn’t know you were ranking for. The pages Google has indexed (and the ones it hasn’t). Core Web Vitals scores from real users. URL inspection results that tell you exactly what Google sees when it crawls your pages.

The problem is that the Search Console interface makes it slow to dig into this data. You can see top-level numbers quickly, but surfacing specific insights — which pages have high impressions but poor CTR, which queries are driving positions 11-20 (the “striking distance” opportunities), which URLs have indexing issues — requires a lot of clicking, filtering, and exporting.

The Google Search Console MCP server solves this by giving Claude direct API access. You ask, Claude queries, you get answers — with the full context of everything Claude already knows about SEO strategy.

What the GSC MCP Server Can Access

The connector gives Claude access to the full Search Console API, which includes search analytics data (queries, pages, devices, countries), URL inspection results, sitemap status, and coverage reports. You can query by any date range, filter by dimension, and compare periods.

Some of the most valuable things you can do: identify every query where your site appears in the top 20 but you’re getting a CTR below average for that position (title/meta optimisation opportunities), find pages with strong impressions but declining clicks (traffic leaks), and spot URLs that Google knows about but isn’t indexing (technical issues to investigate).

Step 1: Set Up OAuth Credentials

The GSC MCP server uses OAuth 2.0 authentication rather than a service account key. Go to the Google Cloud Console, select your project (or create one), and enable the Google Search Console API.

Under APIs & Services → Credentials, create an OAuth 2.0 Client ID. Choose “Desktop App” as the application type. Download the OAuth client JSON file.

Step 2: Install and Configure the GSC MCP Server

The recommended server package for this is @modelcontextprotocol/server-google-search-console — verify the current package name in the MCP server registry as the ecosystem updates regularly. Install it globally with npm, or use npx to run it directly.

Add it to your Claude desktop config (claude_desktop_config.json) alongside your GA4 entry:

{
  "mcpServers": {
    "google-analytics": { ... },
    "google-search-console": {
      "command": "npx",
      "args": ["-y", "@modelcontextprotocol/server-google-search-console"],
      "env": {
        "GSC_OAUTH_CLIENT_FILE": "/Users/yourname/.config/gsc-mcp/client_secret.json",
        "GSC_SITE_URL": "https://www.yoursite.com/"
      }
    }
  }
}

The GSC_SITE_URL must match exactly how your property is registered in Search Console — check whether it uses http or https, with or without www.

Step 3: Authorise Access

The first time Claude tries to use the GSC server, a browser window will open asking you to log in to your Google account and grant access. Authorise it with the account that has access to your Search Console property. The token is saved locally, so you only need to do this once.

Queries That Change How You Do SEO

With both GA4 and GSC connected, Claude can now see both the demand side (what users search for and click) and the behaviour side (what they do when they arrive). This combination is where it gets really interesting:

  • “Find all queries where I rank between position 4 and 15 with more than 100 impressions per month but a CTR below 3%. These are my highest-priority title tag optimisation opportunities.”
  • “Cross-reference my Search Console top pages with GA4 — which pages rank well and get clicks but have a high bounce rate? Those need content or UX work.”
  • “Show me any queries that drove significant clicks six months ago but have dropped off now. I want to understand whether this is a ranking loss or a seasonality issue.”
  • “Which of my pages have impressions growing month-on-month but aren’t converting in GA4? These are pages with organic demand that aren’t doing their job.”

Using URL Inspection at Scale

One underused capability is the URL inspection tool. In the Search Console interface you can only check one URL at a time. Via the API — and therefore via Claude — you can inspect multiple URLs in a single conversation and ask Claude to identify patterns in the results.

For example: “Inspect these 10 URLs that aren’t appearing in Search Console. Tell me which ones are indexed, which are excluded, and what the reason code is for each excluded URL.” Claude will run through them systematically and give you a structured summary.

Next up: Part 4 — Setting Up FreeCrawl (pending), the crawler that gives Claude 96 dedicated SEO tools for deep technical analysis.

SERIES NAVIGATION

  1. Introduction to the AI-Powered SEO & CRO Stack
  2. Setting Up Google Analytics 4 MCP
  3. Setting Up Google Search Console MCP ← You are here
  4. Setting Up FreeCrawl: Your AI-Ready Site Crawler (pending)
  5. Running Multiple Microsoft Clarity Projects (pending)
  6. Powerful SEO Prompts Using Your MCP Stack (pending)
  7. Powerful CRO Prompts Using Your MCP Stack (pending)
  8. Clever SEO Automations with Claude (pending)
  9. Clever CRO Automations with Claude (pending)

Leave a Reply

Your email address will not be published. Required fields are marked *