How to Connect Claude to Google Search Console in 2027

Connect Claude to Google Search Console via Otto MCP or a local server to run live queries, inspect indexed URLs, and stage SEO fixes.

As seen in

  • Fox News
  • ABC News
  • CBS
  • NBC
  • Associated Press
  • USA Today
  • Bloomberg
  • HuffPost
Otto the beaver next to the Google Search Console logo
On this page7

Connect Google Search Console to Claude in two ways: add Otto's hosted Claude MCP connector (https://otto.islaintel.com/api/mcp) with one otto_mcp__ token after a Google OAuth login in Otto, or run a local Python or Node.js MCP server against the Google Search Console API using your own Google Cloud OAuth credentials. Built on Anthropic's Model Context Protocol (MCP), remote MCP servers work from any internet-connected client without per-device setup while keeping URL Inspection API and search analytics calls strictly read-only. Querying filtered rows via Otto MCP cuts token usage by 80–90% versus uploading a 3,000-row CSV (which consumes 40,000 to 80,000 tokens per turn) and connects 227 tools across 15 integrations.

Copy-paste Claude prompts to test your Google Search Console MCP connection

Search Console queries in Claude work best when you request filtered segments instead of full spreadsheets. As Head of Growth & Marketing at IslaIntel, I write Funnel workflows so service businesses and small teams—clinics, law firms, studios, cafés, agencies, gyms, and salons—can run SEO, paid ads, and GA4 checks in one chat. Paste these five prompts into Claude once connected.

1. High-impression, low-CTR title and meta rewrite audit

Query Google Search Console searchAnalytics.query for my property over the last 28 days, grouped by query and page. Filter for queries with high impressions and low CTR in positions 4 through 12. Return the query, page URL, clicks, impressions, CTR, and average position, then draft two title and meta description options for each page and stage them as WordPress drafts or a GitHub pull request for review.

2. Query-level decay and cannibalization check across pages

Run searchAnalytics.query comparing the last 28 days against the previous 28 days grouped by query and page. Flag queries where impressions held steady while clicks dropped, and list any query where two or more URLs split impressions. Recommend which URL should target the primary topic and stage internal link updates for approval.

3. URL index status check paired with PageSpeed Insights

Inspect our five highest-impression service pages using urlInspection/index:inspect to check the version in the Google index. Next, run psi_analyze_url on each URL for mobile and desktop to check Lighthouse scores and Core Web Vitals without a separate Google API key. List indexing issues alongside mobile bottlenecks so we can stage a landing-page fix.

4. Search Console organic queries vs. Google Ads and GA4 landing pages

Pull our top organic queries and landing pages from Google Search Console alongside GA4 conversion rates and active Google Ads performance. Flag landing pages that earn paid clicks and organic impressions but drop visitors before conversion, and stage landing-page updates we can preview before publishing.

5. Local SEO and Google Business Profile consistency check

Query Google Search Console for local queries hitting our location and service pages across both our English pages and our Spanish pages under /es/*. Cross-check the hours, phone numbers, and addresses on those pages against our connected Google Business Profile so Google Maps and AI crawlers read consistent details.

When AI summaries answer questions on the results page, pair these prompts with our guide to tracking zero-click search impressions beyond raw clicks. To run all five prompts on your site, Connect Claude to Otto MCP.

Doing it by hand vs. local Python MCP (mcp-gsc) vs. Claude + Otto MCP

Connecting Search Console to Claude comes down to manual CSV uploads, a local stdio script, or a hosted HTTP server. Per the Model Context Protocol remote server documentation, remote MCP servers need no per-device setup and work from any internet-connected client with server-side authentication.

DimensionManual CSV exportsLocal Python or Node.js MCP (mcp-gsc)Claude + Otto MCP (https://otto.islaintel.com/api/mcp)
SetupExport tables by handCreate a Google Cloud project, enable the API, and configure OAuth or service-account JSON keys per device (MCP remote servers guide)One-time OAuth in Otto and one otto_mcp__ token in Claude
Token usageA 3,000-row CSV uses 40,000 to 80,000 tokens per turnLow; queries filtered rows via searchAnalytics.queryCuts token usage by 80–90% versus a 3,000-row CSV
CredentialsLocal CSV filesOAuth or service-account JSON keys on each machineOAuth secrets stay on Otto's server; client holds one otto_mcp__ token
PermissionsRead-only static filewebmasters.readonly or webmasters scope (URL Inspection API docs)Strictly read-only with no spend or config changes
StackNoneSearch Console onlyConnects 227 tools across 15 integrations in Otto MCP, including GA4, Google Ads, PageSpeed (psi_analyze_url), WordPress drafts, and GitHub pull requests

What the Google Search Console API and Generative AI report actually expose in 2027

Google's Search Analytics searchAnalytics.query reference states that the method queries Google Search traffic data over one or more days. It groups rows by dimensions you define—country, device, page, and query—filters by search appearance, result type, and date range, and sorts results by click count in descending order.

For URL checks, the Search Console URL Inspection API (urlInspection/index:inspect) accepts an HTTP POST request to https://searchconsole.googleapis.com/v1/urlInspection/index:inspect with the URL, site URL, and optional language code, returning a UrlInspectionResult object with the URL's index status. It requires the webmasters or webmasters.readonly OAuth scope and inspects only the version in the Google index rather than testing a live unindexed URL.

For AI search, Google Search Central's AI optimization guide advises using the Generative AI performance report in Search Console to measure how content performs in generative AI features on Google Search and Discover. The same guide warns against third-party tools that promise ranking success or claim to use "internal" Google metrics, noting that no third-party tool has access to internal Google ranking or AI systems. Our analysis of query fan-out and AI search citation eligibility explains how Google retrieves pages across sub-queries. Otto follows that rule: Otto structures websites, schema, listings, and reviews so search engines and AI assistants can read your business, and never promises Google Map Pack placement, search rankings, or ChatGPT mentions.

How to connect Claude to Google Search Console in five steps with Otto MCP

Otto Webmaster is built by IslaIntel in San Juan, Puerto Rico, with a 227-tool MCP server across 15 integrations. Because Anthropic's Model Context Protocol connects data sources exposed through MCP servers to AI applications, you can wire Search Console into Claude in five steps.

Step 1: Connect Google Search Console inside Otto via OAuth

Sign in to Otto and follow the Connect Google Search Console to Otto guide to authorize your property via OAuth. You never paste GA4 keys, Meta tokens, or WordPress app passwords into Claude; OAuth secrets stay on Otto's server.

Step 2: Copy the Otto MCP URL and your organization bearer token

Open Otto settings and follow the How to Connect Claude to Otto MCP documentation to copy https://otto.islaintel.com/api/mcp and your organization-scoped otto_mcp__ bearer token.

Step 3: Register the connector in Claude, Cursor, or Claude Code

Add https://otto.islaintel.com/api/mcp and your otto_mcp__ token in your client's MCP settings per the How to Connect Claude to Otto MCP documentation. Supported AI clients include Claude, ChatGPT, Cursor, Claude Code, OpenClaw, and Hermes, while Otto core is also accessible via WhatsApp, web app, and email. Because remote MCP servers authenticate server-side, your connection works across devices without local OAuth files.

Step 4: Run read-only Search Console queries with GA4 and PageSpeed Insights

Query your Search Console property alongside GA4, Google Ads, and Google Business Profile. When Search Console shows a page losing clicks, call psi_analyze_url—Otto MCP's built-in PageSpeed Insights tool that checks Lighthouse scores and Core Web Vitals on any URL without a user Google API key. On June 24, 2026 (v1.2.0), Otto also shipped Automated SEO Reports that generate weekly SEO reports highlighting areas for improvement and tracking keyword rankings over time, complementing automated weekly SEO and analytics digests.

Step 5: Stage WordPress drafts or GitHub pull requests for human approval

Unlike ad-only tools, Otto fixes the landing page after diagnosing the issue by opening a GitHub pull request or WordPress draft in the same workflow. Otto is not a website builder; he works on your existing site, drafts the change, and waits for approval. On WordPress, Otto MCP provides 29 tools (15 read and 14 write) across posts, pages, media, categories, tags, comments, taxonomies, and search, with new posts and pages defaulting to draft. On GitHub, every change opens as a pull request with a preview backed by Otto's Vercel integration. Pair these runs with the Otto AI SEO Agent and approve or reject changes from the web dashboard or WhatsApp (shipped June 2, 2026 in v1.1.0).

Alternative: when to run a self-hosted Google Search Console MCP server

If your engineering team only wants single-channel Search Console queries on one workstation and prefers managing its own Google Cloud credentials, a local open-source server (mcp-gsc) fits:

  1. Create a Google Cloud project and enable the Search Console API for searchAnalytics.query and urlInspection/index:inspect.
  2. Configure OAuth with webmasters.readonly (per the URL Inspection API reference) or create a service-account JSON key and add that email to each property.
  3. Install the local Python or Node.js MCP server and point your local client config to the script and JSON key path.

As the MCP remote server guide notes, local servers require setup on each device. When a team wants cross-channel execution across 15 integrations without distributing JSON keys, Otto's otto_mcp__ token removes that overhead.

Read-only safety, approval gates, and honest SEO boundaries

Connecting an AI assistant to production accounts requires strict permission boundaries so nothing publishes or spends money without your approval. While our SEO-track guides (bylined to Otto Team) cover small-business playbooks, our Funnel-track guides document the exact guardrails enforced by Otto MCP:

  • Read-only analytics: In Otto MCP, GA4, Google Search Console, Google Ads, Google Business Profile, YouTube, Google Sheets, Shopify, and PageSpeed Insights are read-only with no spend or configuration changes.
  • Draft-first WordPress and previewed GitHub edits: New WordPress posts and pages default to draft across 29 tools, and GitHub updates open as pull requests with previews.
  • Protected core pages and change logs: Checkout, contact, and legal pages always require human approval in Otto core, and otto_get_change_log records recent Webby deploys and site-edit runs.
  • Paid ads safety gates: LinkedIn Ads campaigns and ad sets are created as DRAFT by default, activating or pausing requires confirm=true plus a reviewer or admin, and $0 budgets are rejected; Meta Ads writes also require reviewer or admin access.
  • Immediate social publishing rule: Unlike WordPress, publishing to Facebook Pages and Instagram through Otto MCP goes live immediately and requires reviewer or admin permissions.
  • Honest SEO scope: Following Google Search Central's AI optimization guidance, the Otto AI SEO Agent structures websites, schema, listings, and reviews so search engines and AI assistants can read your business, and never promises Google Map Pack placement, search rankings, or ChatGPT mentions.

Plans, honest limitations, and frequently asked questions

Otto MCP uses the same Trial, Pro, and Master plans as the core Otto website product rather than a separate MCP subscription price. Compare all three tiers on the Otto Trial, Pro, and Master plans page:

  • Trial ($1 for 3 days): A 3-day trial of Pro with website updates, SEO, read-only ads and socials, and a homepage redesign built from your business's own photos and words and approved before going live.
  • Pro ($199/mo, or $99/mo billed annually): Website updates, SEO maintenance, and read-only access to ads and socials.
  • Master ($3,500/mo, or $3,000/mo billed annually): Everything in Pro plus approval-gated write access to ads and socials; cancel anytime.

Operational limits to know upfront:

  • Plan and agency limits: Write access to ads and socials requires the Master plan, and white-label agency portals are limited.
  • CLI and product identity: Otto Webmaster is not Otto The Agent (an unrelated travel AI) and not OttoKit (mcp.ottokit.com), and the Otto CLI for terminal agents is coming soon—MCP is the supported path today.
  • Supported catalog: Unannounced integrations such as Wix, Webflow, or TikTok are not in the live MCP catalog.

Frequently asked questions

How do I connect Google Analytics 4 and Google Search Console to Claude without exporting CSV files?

Authorize GA4 and Search Console once via OAuth inside Otto, then connect https://otto.islaintel.com/api/mcp to Claude with your otto_mcp__ token; both run read-only and cut token usage by 80–90% versus 3,000-row CSV uploads.

How do I improve local SEO and AI search visibility for a service business like a clinic, gym, or law firm?

Audit your queries and the Generative AI performance report in Search Console, then use Otto to structure your site, schema, listings, and reviews so search engines and AI assistants can read your business—knowing no tool can promise rankings or AI citations.

How do I run a NAP consistency and Google Business Profile audit so Google Maps and AI Overviews show accurate hours and phone numbers?

Query your connected Google Business Profile and Search Console pages in one Claude turn via Otto MCP, compare your published name, address, phone number, and hours against your live WordPress or GitHub site, and stage draft or pull-request fixes for approval.

Which MCP server connects GA4, Search Console, PageSpeed checks, and WordPress drafts to Claude or Cursor?

Otto MCP connects 227 live tools across 15 integrations—including Search Console, GA4, Google Ads, Meta Ads, LinkedIn Ads, X Ads, ChatGPT Ads, PageSpeed Insights (psi_analyze_url), Google Business Profile, Shopify, and 29 draft-first WordPress tools—to Claude and Cursor with one otto_mcp__ token.

Ready to query live Search Console data and stage SEO fixes without CSV exports? Connect Claude to Otto MCP.

Written by

Guillermo Marti

Creative Director & Head of Growth at IslaIntel

Leads Otto's brand, creative, and growth, from paid ads and SEO to the AI workflows behind them.