Magnifying glass over financial graphs, illustrating Search Console data inspected through an AI assistant

Google Search Console MCP: Connect Claude or ChatGPT to Search Console

There is no official Google Search Console MCP server. Google ships MCP servers for Google Analytics and Google Ads, but as of September 2026 Search Console has none, so the choices are a community server you host with a service-account key, or a hosted MCP that already speaks Search Console. datavessel is the second kind: sign in with Google once, add one connector URL to Claude or ChatGPT, and your assistant can pull top queries and pages, rising terms, device and country splits, keyword opportunities, sitemaps and URL inspection, and hand the findings to agents that fix them. The same connector carries your store; that setup is in connect Claude to your store.

The three ways to get Search Console into an AI assistant

1. Export and paste. Download the performance report, drop the CSV into the chat. Works once. It is stale by tomorrow and blind to anything you did not export.

2. A community MCP server. Several open-source projects wrap the Search Console API. They work, and they suit a developer who is comfortable creating a Google Cloud project, generating a service-account JSON key, granting it access to the property, and running a local server. The key sits on your machine, the server needs maintaining, and it covers Search Console alone.

3. A hosted MCP with Search Console built in. You authorize with Google’s normal sign-in, no service account and no JSON file, and the tools appear in whatever assistant you use. That is the route below.

What Claude or ChatGPT can do once connected

datavessel exposes eleven Search Console tools, all reads, so nothing here can change anything in Google:

  • Top queries and top pages for any window: clicks, impressions, CTR, position.
  • Rising queries, the terms gaining impressions week over week, which is where new content and retitles come from.
  • Keyword opportunities: queries with impressions and a position between 5 and 20, ranked by what a CTR or position gain would be worth.
  • Performance by device and by country.
  • URL inspection: is this page indexed, when was it last crawled, does Google agree with the canonical.
  • Sitemaps: which are submitted, when they were last read, errors and warnings.

In practice the questions look like this:

  • “Which queries gained impressions in the last 14 days that no page of mine targets?”
  • “List pages sitting at positions 8 to 15 with more than 100 impressions, with their titles”
  • “Has Google re-crawled these four URLs since I rewrote them?”
  • “Compare mobile and desktop CTR for my collection pages”
  • “Which product pages lost clicks month over month, and which queries dropped?”

Because the store, GA4, Google Ads and Merchant Center are in the same connector, the follow-ups work too: “of those losing pages, which products are low on stock or disapproved in Merchant Center.”

Connect Claude to Search Console: step by step

  1. Sign in at app.datavessel.io and add Google Search Console as a source. It is a standard Google sign-in; pick the properties to expose. You can revoke it from your Google account at any time.
  2. Add the connector to Claude. In Claude Desktop, Settings, then Connectors, add a custom connector with https://mcp.datavessel.io/mcp and sign in. In Claude Code, add the same URL as an MCP server.
  3. Ask. “Which Search Console properties can you see?” then the first real question.

Connect ChatGPT to Search Console

Same server, different door. In the ChatGPT desktop app enable developer mode under Settings, then Connectors, add a custom connector with the same URL, sign in, and the Search Console tools appear next to your other connectors. The walkthrough with screenshots is the same as for the store: connect ChatGPT to Shopify, steps two and three.

From questions to fixes

Reading Search Console is the first half. The reason to put it behind the same connector as your store is the second half: the pages the data points at are the ones the assistant can also change. Thin category descriptions, missing meta titles, products whose queries are slipping, all reachable through the store tools, and every change proposed for your approval before it lands.

Then take it off the chat. Inside datavessel the same tools run on a schedule: a weekly SEO audit that pulls the opportunity list and executes the fixes it is allowed to, and an AEO harness that pairs Search Console with citation scans across ChatGPT and Claude. The terminal version of the loop is in AEO from the command line.

Search Console MCP compared

Export and paste Community MCP server datavessel
Setup None Cloud project, service account, JSON key, local server Google sign-in, one URL
Freshness Stale on export Live Live
Works in Any chat Claude Desktop, Claude Code, Cursor Claude, ChatGPT, Cursor, Gemini CLI, plus CLI and mobile
Covers What you exported Search Console Search Console plus store, GA4, ads, Merchant Center, and 13 more
Can fix what it finds No No Yes, through the store tools, with approval
Runs on a schedule No No Yes
Secrets None On your machine OAuth grant, revocable
Cost Free Free plus your time Free for one store

Frequently asked questions

Is there an official Google Search Console MCP server?

Not as of September 2026. Google publishes official MCP servers for Google Analytics and Google Ads. Search Console is covered by community servers and by hosted platforms such as datavessel that include it.

Do I need a service account to connect Search Console to Claude?

Only with a community server. datavessel uses Google’s standard OAuth sign-in, so there is no Cloud project, no service-account JSON, and nothing to rotate. Access is revocable from your Google account.

Can the assistant change anything in Search Console?

No. All eleven tools are reads. The writes happen on the store side, where the pages live, and every one of those waits for your approval.

Does it work with Google Analytics too?

Yes. GA4 is a separate source in the same connector with its own tools, including standard and realtime reports, so “traffic dropped on these pages, which queries lost impressions” is one conversation.

Can I do this from the terminal instead of a chat?

Yes. The datavessel CLI runs the same tools with JSON output, which is how the monthly measurement loop in the e-commerce CLI is built.

What does it cost?

The connector is included in every plan, including the free one for a single store. Model usage runs through your own Claude or ChatGPT subscription.

The bottom line

Search Console is the one Google product without an official MCP server, and the one whose data most directly tells a store what to fix next. Put it behind the same connector as the store, and the assistant that finds the slipping page is the one that proposes the fix.

Search Console, GA4 and the store in one connector, every fix behind your approval: the datavessel MCP server →


Posted

in

by

Tags:

Comments

Leave a Reply

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