Claude SEO Audit: GA4 & GSC MCP Guide - 12 Copy-Paste Prompts

Claude SEO Audit: GA4 & GSC MCP Setup Guide

Connect Google Search Console and Google Analytics 4 directly to Claude via MCP (Model Context Protocol). Run real-time SEO audits, surface quick wins, and get AI-powered recommendations without leaving your terminal or IDE.

You'll set up two separate MCP servers — one for Google Search Console (GSC) and one for Google Analytics 4 — then wire both into your MCP-compatible client (e.g. Claude Desktop) plus a shared Google Cloud project for auth.

Below is a concise walkthrough assuming Claude Desktop, but the same pattern works for any MCP client.

Overview: What You Need

• A Google Cloud project with:
  • Search Console API enabled (for GSC MCP)
  • Analytics Data API enabled (for GA4 MCP)
• OAuth credentials (Desktop app) for both MCP servers
• Your:
  • GSC property name (often sc-domain:yourdomain.com)
  • GA4 property ID(s)

Step 1: Create / Use a Google Cloud Project

1. Go to Google Cloud Console and create a new project (e.g. "SEO Analytics MCP").
2. In "APIs & Services > Library", enable:
   • "Google Search Console API"
   • "Analytics Data API" (for GA4)

This gives both MCP servers a single project to authenticate against.

Step 2: Create OAuth Credentials (Desktop App)

1. In the same project, go to "APIs & Services > Credentials".
2. If prompted, configure an External OAuth consent screen (name like "MCP SEO Analytics").
3. Click "Create credentials > OAuth client ID > Application type: Desktop app".
4. Name it (e.g. "MCP Desktop OAuth") and click Create, then download the JSON file.

You'll reuse this same OAuth client for both GSC and GA4 MCP, unless your server docs say otherwise.

Step 3: Install / Point to the GSC MCP Server

Most current guides use a ready-made GSC MCP server you call via npx mcp-remote (similar to the GA4 one).

In Claude Desktop:

1. Go to Settings > Developer > Edit Config.
2. In the JSON, under mcpServers, add a block for GSC, roughly:

{
  "mcpServers": {
    "gsc-mcp": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-remote@0.1.30",
        "https://<gsc-mcp-host>/mcp"
      ],
      "env": {
        "GOOGLE_OAUTH_CLIENT_JSON": "/full/path/to/your_oauth_client.json",
        "GSC_SITE_URL": "sc-domain:yourdomain.com"
      }
    }
  }
}

Key details:

• Use sc-domain:yourdomain.com if your property is a Domain property; https://yourdomain.com/ will fail in that case.
• GOOGLE_OAUTH_CLIENT_JSON points to the OAuth JSON you downloaded.

Restart Claude; on first use, you'll be prompted to grant GSC access, and once complete the server should show as "Running".

Step 4: Install / Point to the GA4 MCP Server

A typical GA4 MCP server is also exposed via mcp-remote, for example https://mcp-ga.stape.ai/mcp in one popular setup.

In the same Claude config, add another block:

{
  "mcpServers": {
    "gsc-mcp": {},
    "ga4-mcp-server": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-remote@0.1.30",
        "https://mcp-ga.stape.ai/mcp"
      ],
      "env": {
        "GOOGLE_OAUTH_CLIENT_JSON": "/full/path/to/your_oauth_client.json",
        "GA4_PROPERTY_ID": "123456789"
      }
    }
  }
}

Notes:

• GA4_PROPERTY_ID is your GA4 property ID from the GA4 Admin UI.
• As with GSC, you'll authenticate once via a browser window and then the MCP will appear as "Running" in Settings > Developer.

Step 5: Confirm Access and Start Using

Once both are running, you can:

• Ask things like "Show me last 28 days branded query clicks for mysite.com" via the GSC server.
• Ask "Compare organic vs paid conversions by landing page in GA4 for last month" via the GA4 server.

Some GSC MCP servers expose tools for queries, indexing checks, sitemap management, and URL inspection.

Minimal Checklist

• One Google Cloud project with Search Console API + Analytics Data API enabled
• OAuth Desktop client JSON downloaded and stored locally
• GSC MCP config with GSC_SITE_URL = sc-domain:yourdomain.com (or exact URL property)
• GA4 MCP config with correct GA4_PROPERTY_ID
• Both appear "Running" in Claude's Developer settings after auth

Google Search Console Prompts

Copy-paste these prompts directly into Claude once your GSC MCP is connected.

Prompt 1: Quick Health Check

Pull the last 28 days of GSC data. Show me:

• Total clicks, impressions, avg CTR, avg position
• Top 20 queries by clicks
• Any queries where position is between 4-10 (quick win zone)
• Pages with impressions > 1000 but CTR < 2%

Prompt 2: Keyword Cannibalization Audit

Check for keyword cannibalization:

• Find queries where 2+ pages rank for the same keyword
• Sort by impressions (highest first)
• For each cannibalized query, recommend which page should be the primary target
• Suggest redirects or content consolidation where appropriate

Prompt 3: Content Decay Detection

Compare the last 28 days vs the previous 28 days:

• Pages with >20% drop in clicks
• Pages with >3 position drop on their top queries
• Queries that lost impressions significantly

Group by severity: critical (>50% drop), warning (20-50%), watch (10-20%)

Prompt 4: Quick Wins Report

Find my highest-impact quick wins:

• Queries at positions 8-20 with high impressions
• Calculate estimated click gain if moved to position 3
• Sort by potential traffic gain
• For each, suggest whether to optimize existing page or create new content

Prompt 5: New vs Lost Keywords

Compare this month vs last month:

• New queries that appeared (didn't exist last month)
• Lost queries that disappeared
• Queries with biggest position improvements
• Queries with biggest position declines

Summarize trends and recommend action items.

Prompt 6: Page Performance Audit

Audit my top 50 pages by impressions:

• Current clicks, impressions, CTR, position for each
• Flag pages with CTR below expected for their position
• Identify pages that rank for high-volume queries but underperform
• Recommend title tag and meta description improvements for bottom 10 performers

Google Analytics (GA4) Prompts

These prompts work with the GA4 MCP server connected.

Prompt 1: Traffic Overview

Pull GA4 data for the last 30 days:

• Total sessions, users, new users, engaged sessions
• Traffic by channel (organic, direct, referral, social, paid)
• Top 20 landing pages by sessions
• Engagement rate and avg session duration by channel

Prompt 2: Conversion Funnel Analysis

Analyze the conversion funnel:

• Landing page sessions to key event (sign up, purchase, contact form)
• Conversion rate by traffic source
• Conversion rate by landing page (top 20)
• Drop-off points between page views and conversions

Identify the highest-converting and lowest-converting paths.

Prompt 3: Content Engagement Scoring

Score my content by engagement:

• For each of my top 50 pages, pull: sessions, engaged sessions, avg engagement time, bounce rate, events
• Create an engagement score (weighted: engagement time 40%, engaged session rate 30%, events per session 30%)
• Rank pages by engagement score
• Flag high-traffic but low-engagement pages (need improvement)
• Flag low-traffic but high-engagement pages (need promotion)

Prompt 4: User Journey Mapping

Map the most common user journeys:

• Top 10 landing page > second page paths
• Where do users go after reading blog posts?
• What percentage of blog readers visit product/pricing pages?
• Which blog topics drive the most downstream conversions?

Prompt 5: Geographic & Device Analysis

Break down performance by:

• Top 15 countries by sessions
• Device category (desktop, mobile, tablet) with engagement rate for each
• Conversion rate by device type
• Any geographic regions with unusually high or low engagement

Recommend targeting or optimization priorities.

Prompt 6: Revenue Attribution (Ecommerce)

If ecommerce tracking is enabled:

• Revenue by traffic channel
• Revenue by landing page
• Average order value by source
• Top converting product pages
• Revenue trend (weekly) for last 12 weeks

Identify which content directly drives revenue.

Combined Analysis: GSC + GA4

The real power is combining both data sources. Use these after connecting both MCPs.

Traffic-to-Conversion Analysis

Cross-reference GSC and GA4 data:

• For my top 30 organic landing pages (from GSC), pull GA4 engagement and conversion data
• Find pages with high organic traffic but low engagement (content mismatch)
• Find pages with high engagement but low organic traffic (SEO opportunity)
• Recommend which pages to prioritize for optimization

SEO Gap Analysis

Using both GSC and GA4:

• Pull pages that get organic impressions but low clicks (CTR problem)
• Cross-reference with GA4 to check if those pages convert well when visited
• If high conversion + low CTR: priority title/meta optimization
• If low conversion + low CTR: consider content rewrite
• Create a prioritized action list with estimated impact

Geographic Expansion Opportunities

Combine GSC and GA4 geo data:

• GSC: which countries show impressions but low clicks?
• GA4: which countries have highest engagement/conversion rate?
• Find overlap: high-engagement countries where we're not ranking well
• Recommend content localization or targeting priorities

Revenue-Driven SEO Priorities

Connect organic performance to revenue:

• GSC: top organic queries driving traffic
• GA4: which of those landing pages generate revenue?
• Calculate estimated revenue per organic keyword
• Prioritize SEO efforts by revenue impact, not just traffic volume
• Identify high-revenue keywords where we could improve position

Advanced Workflows

Weekly SEO Audit Automation

Run this every Monday morning:

Weekly SEO check:

1. GSC: Compare last 7 days to previous 7 days (clicks, impressions, CTR, position)
2. Flag any pages with >10% click drop
3. GA4: Check if organic engagement rate changed
4. List new keywords that appeared this week
5. List keywords that dropped out of top 20
6. Summarize in a table and recommend top 3 actions for this week

Monthly Content ROI Report

Monthly content performance:

1. All content published this month
2. GSC: impressions and clicks for each
3. GA4: engagement and conversions for each
4. Calculate content ROI score (traffic + engagement + conversions)
5. Compare to previous month's published content
6. Recommend content topics for next month based on what's working

Tips and Best Practices

• Start with GSC if you only set up one. Search Console data is the most actionable for SEO.
• Date ranges matter. Always specify explicit date ranges to get consistent, comparable data.
• Combine with your knowledge base. Tell Claude about your target keywords, business goals, and competitors for more relevant recommendations.
• Save your best prompts. Turn them into Claude Code skills or slash commands for one-click audits.
• Run weekly. The compounding value of regular audits is massive. Set a Monday morning routine.

Troubleshooting Common Issues

OAuth Token Expired

If you get authentication errors after a period of inactivity, delete the cached token file and re-authenticate. The token file location varies by MCP server implementation - check the server's docs.

Wrong Property Type

GSC has two property types: Domain properties (sc-domain:example.com) and URL-prefix properties (https://example.com/). Using the wrong format is the most common setup error. Check your GSC dashboard to confirm which type you have.

GA4 Property ID Not Found

The GA4 property ID is a number (e.g., 123456789), not the measurement ID (which starts with G-). Find it in GA4 Admin > Property Settings.

MCP Server Not Showing as Running

Make sure you've saved the config file and fully restarted Claude Desktop. Check that npx is available in your PATH and that Node.js is installed.

Resources

• AI Topia Community (free)
• Book a strategy call