The
GSC
+
GA4
MCP
for Claude and ChatGPT
- Job: Chat with LLMs to understand your SEO performance, why it’s changing, and what to do next.
- Problem: Most MCPs pipe in raw data, so the model works harder and hallucinates more.
- Solution: Our MCP filters data first and reads and writes annotations, so you know why things happen.
Get Started for Free
Read-only access to Search Console. Disconnect anytime.
Already on SEO Gets? Connect it nowSix data sources behind one connection
-
Google Search Console
Clicks, impressions, CTR, and position by query, page, country, and device.
-
Google Analytics 4
Sessions, engagement, key events, and revenue, joined to the same pages and dates.
-
Google Discover
Discover clicks, impressions, and CTR by page, country, and content group.
-
Indexing status
Coverage state and last crawl date for each page, plus the pages at risk of dropping out.
-
Content changes
Detected edits to copy, titles, headings, and links, plus HTTP status changes.
-
Annotations
Your notes on what shipped and when, next to Google core and spam updates.
8 reasons to choose our MCP
-
One-click setup
Connect your Google account to SEO Gets with one click, then add us from Claude’s or ChatGPT’s connector library. No config files, no API keys, no developer.
-
It reads and writes
Most GSC and GA4 connectors are read-only. Ours lets your LLM pull performance data, then log annotations, create content groups, and update topic clusters in the same conversation.
-
Annotations that build institutional memory
Ask your LLM to annotate a title tag change, a migration, or a homepage H1 update right when it happens. Six months later, you know exactly why traffic moved, and so does everyone on your team.
-
Content grouping and prioritization
Your LLM sees performance by content group and topic cluster, not a flat list of 50,000 URLs. It can tell you which section of the site deserves attention first.
-
Pre-filtered data, fewer hallucinations
Raw GSC exports force the model to wade through huge tables and guess. We filter and aggregate before the data reaches the model, so the answers are grounded in real numbers.
-
Six data sources, one connection
GSC, GA4, indexing status, annotations, content changes, and Discover all sit behind a single MCP.
-
Every site in one conversation
Ask which of your sites lost the most clicks this month, or group client sites into a portfolio and compare them. No switching properties, no stitching exports together.
-
Built by SEOs who manage real sites
We use this every day across our own properties and client sites. Every tool in it exists because we needed it for actual work.
Matthew Mellinger
Co-founder of SEO Gets
Why you can trust the answers
Most connectors hand the model a raw export and hope. We do the filtering first.
A typical GSC or GA4 connector
- You ask a question
- The connector pulls raw rows from the API
- Thousands of rows land in the model’s context
- The model skims, samples, and guesses
- Made-up trends and numbers
- Read-only, so nothing gets logged
- One data source per connector
The SEO Gets MCP
- You ask a question
- SEO Gets filters and aggregates your data first
- The model gets a short, exact table, by content group and topic cluster
- You get an answer you can check against your dashboard
- Answers grounded in real numbers
- Reads and writes in the same conversation
- GSC, GA4, indexing, Discover, and your change history together
What SEOs use it for
get_site_performancelist_content_changes More prompts to try
Copy one, swap in your site, and go.
- “Rank my sites by clicks lost in the last 28 days.”
- “What changed on /pricing before clicks dropped?”
- “Compare this March to last March for the Blog group.”
- “How did Discover clicks move after the last core update?”
- “Which topic cluster grew the most this year?”
- “Create a content group for every URL under /docs/.”
“SEO Gets built that MCP connector. Genuinely one of my favorite products right now.”
Sarthak Ahuja SEO @ Favikon Connected in about two minutes
No config files, no API keys, no developer.
-
Connect your Google account
Sign in to SEO Gets with Google. One click, and your Search Console and GA4 properties are in. Already on SEO Gets? Skip to step 2.
-
Add SEO Gets in Claude or ChatGPT
We’re an approved connector in both libraries. Open our listing in the one you use and click Connect.
-
Ask your first question
Approve access when your LLM asks, then start with something simple. It gets better as you add content groups and annotations.
Show me top pages for example.com last month.
On another MCP client? Paste this URL:
https://app.seogets.com/mcp Every tool your LLM can call
28 tools. 13 read your data, and 15 write to SEO Gets when you ask them to.
Performance
-
get_site_performanceRead - Search Console and GA4 data for a site, joined on the same pages and dates. Groups by query, page, country, device, content group, or topic cluster, and splits branded from non-branded.
-
list_site_queriesRead - Every search query a site ranked for in the last 16 months.
-
list_site_pagesRead - Every page that got search traffic in the last 16 months.
-
get_discover_performanceRead - Google Discover clicks, impressions, and CTR by date, page, country, or content group.
-
list_sitesRead - The sites in your account, or only the ones you marked as priority.
Annotations and content changes
-
list_content_changesRead - One timeline of your annotations, detected content changes (copy, titles, meta, headings, links), HTTP status changes, internal link changes, and Google updates.
-
create_site_annotationWrite - Adds an annotation to a whole site, a content group, or specific URLs.
-
update_site_annotationWrite - Edits an annotation’s date, title, description, or scope.
-
delete_site_annotationWrite - Deletes an annotation.
Indexing
-
get_indexing_overviewRead - Total pages, a breakdown by coverage state, and the pages at risk of de-indexing because Google hasn’t crawled them in 90 days or more.
-
get_indexing_statusRead - Coverage state, indexing verdict, last crawl date, and rich results for each page, filtered by URL or status.
Site setup
-
get_property_setupRead - Checks whether a site has its brand summary, branded keywords, content groups, and topic clusters, and which ones are prioritized.
-
setup_propertyWrite - Runs One Click Setup to create whatever is missing. Uses your team’s AI credits.
Content groups and topic clusters
-
list_content_groupsRead - A site’s content groups and the URL filters that define them.
-
create_content_groupWrite - Creates a content group from URL filters, and can mark it as a priority.
-
update_content_groupWrite - Renames a content group, or changes its filters or priority.
-
delete_content_groupWrite - Deletes a content group.
-
list_topic_clustersRead - A site’s topic clusters and the keyword filters that define them.
-
create_topic_clusterWrite - Creates a topic cluster from keyword filters, and can mark it as a priority.
-
update_topic_clusterWrite - Renames a topic cluster, or changes its filters or priority.
-
delete_topic_clusterWrite - Deletes a topic cluster.
-
prioritize_property_groupsWrite - Ranks a site’s content groups and topic clusters by importance to the business, so answers lead with them. Uses AI credits and only runs when you agree.
Portfolios
-
list_portfoliosRead - Your portfolios and site migrations, with the sites, content groups, and topic clusters in each.
-
get_portfolio_performanceRead - Combined Search Console and GA4 data across a portfolio of up to 20 sites, or one row per site to compare them.
-
create_portfolioWrite - Groups sites into a portfolio, or pairs an old and a new site as a migration.
-
add_portfolio_sitesWrite - Adds sites to a portfolio.
-
remove_portfolio_sitesWrite - Removes sites from a portfolio. The sites and their data stay.
-
delete_portfolioWrite - Deletes a portfolio or migration. The sites aren’t touched.
Frequently
Asked
Questions
What is a Google Search Console MCP?
MCP (Model Context Protocol) is the open standard LLMs like Claude and ChatGPT use to connect to other apps. A Search Console MCP lets your LLM query your real GSC data instead of guessing. The SEO Gets MCP puts GA4, indexing, annotations, content changes, and Discover behind the same connection.
How is it different from other Search Console MCPs?
Three ways. It filters and aggregates your data before the model sees it, so answers come from real numbers and not a skim of a raw export. It writes as well as reads: annotations, content groups, and topic clusters. And setup is one click, with no config files or API keys.
Which LLMs does it work with?
Claude and ChatGPT, where SEO Gets is an approved connector in the library. Any other tool that supports MCP can connect with the server URL https://app.seogets.com/mcp.
Do I need a paid plan?
MCP access covers your core sites and Super Sites. Core starts at $49/month, and you can try it free. Free sites aren't available through the MCP, and indexing and content change data come with the Super Site add-on.
Can my LLM change anything?
Only inside SEO Gets, and only when you ask. It can create, edit, or delete annotations, content groups, topic clusters, and portfolios. It can't change anything in Search Console, because SEO Gets only has read-only access there. Claude and ChatGPT also let you turn individual tools off in the connector settings.
Can my whole team use it?
Yes. In team workspaces, an admin can add SEO Gets once. Then each teammate connects and approves access for their own account.
Is my data private?
Yes. SEO Gets is private by design and never sells your data. Each person approves access through a consent flow before anything connects, and you can disconnect anytime.