Google Search Console MCP servers: tools, auth and the indexing gap
A Google Search Console MCP server lets an AI assistant such as Claude read your Search Console data through MCP (Model Context Protocol: a standard way for AI assistants to use a tool directly). Google doesn't offer one. Community servers wrap the Search Console API: they read performance, inspect URLs and submit sitemaps, but the API can't request indexing.
That matters because an agent that finds a page stuck in "Discovered - currently not indexed" still needs another way to act on it. This page compares eight Search Console MCP servers as of October 2026: the tools each exposes, how it signs in to Google, what it costs, and where an indexing tool fits next to them.
How do Search Console MCP servers compare?
The main differences are where the server runs, how it gets access to your Google account, and whether it can change anything. The table covers the servers described below, as published on each project's own page or repository, as of October 2026 (checked on October 17, 2026).
Server
Runs
Tools (as published)
Google sign-in
Can it change anything?
Price
Ahrefs MCP (gsc tool group)
Hosted by Ahrefs
Search Console keywords, pages, positions and history, plus Ahrefs data
Through your Ahrefs account
Not published for the gsc group
Ahrefs Lite plan and up (from $129/mo)
gsc-indexer-mcp (Mrshahidali420)
Your machine (Python)
15 tools, including browser-driven Request Indexing
OAuth with your own Google Cloud project
Yes: sitemaps, Request Indexing
Free (FSL-1.1-ALv2 license)
gsc-mcp (mikusnuz)
Your machine (Node)
Search Console API plus Indexing API tools
OAuth refresh token or service account
Yes: sites, sitemaps, Indexing API
Free (MIT)
mcp-gsc (AminForou)
Your machine (Python)
20 tools
OAuth or service account
Sitemaps; site add/delete off by default
Free (MIT)
mcp-server-gsc (ahonn)
Your machine (Node)
8 tools
Service account
Sitemap submit
Free (README says MIT)
QuerySail
Hosted
5 read-only tools
Sign in once with OAuth
No
$8/mo after a 7-day trial
search-console-mcp (saurabhsharma2u)
Your machine (Node)
Search Console, Bing Webmaster Tools, GA4 and AdSense
OAuth or service account
Yes: sites, sitemaps, submissions
Free (MIT)
search-console-mcp-server (sudomichael)
Your machine (Node)
6 tools and 5 prompts
OAuth, read-only scope
No
Free (MIT)
FreshlyCited (not a Search Console server)
Hosted
Index checks, smart push, IndexNow, client reports
No Google credentials; OAuth 2.1 or API key to FreshlyCited
Pushes spend credits
MCP on Agency Pro ($199/mo) and Agency Scale ($599/mo)
Projects change quickly. Check each repository's README before you install, because tool lists in this space often change between releases.
What can the Search Console API do through MCP?
Any Search Console MCP server can do only what the Search Console API allows, because that's what it calls. Google's API reference lists four resources: Search Analytics ("Query your search traffic data with filters and parameters that you define"), Sitemaps (list, get, submit, delete), Sites (add, delete, get, list) and URL Inspection (Google Search Console API reference). There is no method for requesting indexing.
URL Inspection is the method most people want an agent for, and it's a read. Google describes it as "View the indexed, or indexable, status of the provided URL" and adds: "Presently only the status of the version in the Google index is available; you cannot test the indexability of a live URL" (Google). It accepts either the full webmasters scope or webmasters.readonly, so a read-only server can still inspect URLs.
Google gave the same per-property figures when it launched the API: "2,000 queries per day" and "600 queries per minute" (Google Search Central Blog, January 2022). For a large site, 2,000 inspections a day means an agent can't audit every URL at once; it has to sample, or work property by property.
The Request indexing button lives only in the Search Console interface. Google's help page says "There is a daily limit to how many index requests you can submit" and "Submitting a request does not guarantee that the page will appear in the Google Index" (Search Console Help). Our guide to free ways to submit a URL to Google covers the button, sitemaps and the rest by hand.
So a Search Console MCP server is a very good way to ask questions like "which pages lost clicks this month?" or "is this URL indexed, and which canonical did Google pick?". It isn't a way to get a page indexed. If your agent turns up URLs Google hasn't indexed, check those pages and push the invisible ones.
Which open-source Search Console MCP servers can you run yourself?
Self-hosted servers run on your own machine, usually through npx (Node) or uvx (Python), and need either a service account or OAuth credentials from a Google Cloud project you create. They're free. They're listed alphabetically by GitHub owner, not ranked.
01
mcp-server-gsc (ahonn)
mcp-server-gsc is a Node server. Its source registers eight tools: list_sites, search_analytics, enhanced_search_analytics, detect_quick_wins, index_inspect, list_sitemaps, get_sitemap and submit_sitemap (src/server.ts). The README says it can "Retrieve up to 25,000 rows of performance data" and supports regex filters.
Auth: a service account. You enable the Search Console API in Google Cloud, create a service account key, then add the service account's email to Search Console.
Install:npx -y mcp-server-gsc with GOOGLE_APPLICATION_CREDENTIALS pointing at the key file.
Limits: no OAuth option, so you need Google Cloud access to set it up.
02
mcp-gsc (AminForou)
mcp-gsc is a Python server published as mcp-search-console. Its README lists tools including list_properties, get_search_analytics, compare_search_periods, inspect_url_enhanced, batch_url_inspection ("Inspect up to 10 URLs at once"), check_indexing_issues and manage_sitemaps, and says to call get_capabilities"for the full list of all 20 tools".
Auth: OAuth with your own Google account (recommended in the README) or a service account.
Install:uvx mcp-search-console with GSC_OAUTH_CLIENT_SECRETS_FILE or GSC_CREDENTIALS_PATH set.
Safety:"By default, add_site, delete_site, and delete_sitemap are disabled"; you turn them on with GSC_ALLOW_DESTRUCTIVE.
Limits: the README notes the server requires Python 3.11 or newer if you clone it rather than use uvx.
03
gsc-mcp (mikusnuz)
gsc-mcp covers the Search Console API endpoints (sites, sitemaps, Search Analytics including the hour dimension, URL Inspection) and adds three Indexing API tools: indexing_publish, indexing_get_metadata and indexing_batch_publish.
Auth: an OAuth refresh token (scopes webmasters and indexing) or a service account.
Install:npx -y @mikusnuz/gsc-mcp.
Limits: the README states the Indexing API rule plainly: "Google supports this API only for pages with JobPosting structured data or livestream pages with BroadcastEvent". For any other page, those three tools are outside what Google supports.
04
search-console-mcp (saurabhsharma2u)
search-console-mcp puts "Google Search Console + Bing Webmaster Tools + GA4 + AdSense" behind one set of tools. Its tool table includes sites_list, analytics_query, analytics_compare, inspection_inspect (Google URL inspection and Bing URL info), pagespeed_analyze, seo_audit and indexing_submit, which can send URLs through IndexNow (an open "this page changed" notice that Bing and other participating engines accept; Google doesn't use it).
Auth: OAuth (npx search-console-mcp setup) or a service account for Google; an API key for Bing Webmaster Tools.
Install:npx search-console-mcp after setup; there's also a one-click .mcpb bundle for Claude Desktop.
Limits: it can add and remove properties and sitemaps, so give it the narrowest access you can.
05
search-console-mcp-server (sudomichael)
search-console-mcp-server aims to skip the Google Cloud setup. Its tools are list_properties, query_search_analytics, compare_periods, inspect_url, inspect_urls (up to 10 URLs) and list_sitemaps, plus five built-in prompts such as traffic_drop and indexing_audit.
Auth: run npx search-console-mcp-server login and sign in with Google. It requests webmasters.readonly, which the README says "cannot modify your properties, submit sitemaps, or change anything".
Limits: read-only by design, so no sitemap submission.
Which hosted Search Console MCP servers need no Google Cloud project?
Hosted servers run on the vendor's infrastructure. You add a URL to your AI client, sign in, and connect Google once; there's no key file to manage. They cost money.
06
Ahrefs MCP (Search Console tool group)
Ahrefs runs a remote MCP server at https://api.ahrefs.com/mcp/mcp for "users with a Lite plan or higher" (Ahrefs docs). One of its tool groups is gsc: "Google Search Console — keywords, pages, positions, and performance history", and you can load just the groups you need with a tools= parameter (Ahrefs: tool categories).
Auth: an authorization consent screen creates a dedicated key tagged with MCP scope; you can also generate an MCP key manually.
Install (Claude Code):claude mcp add ahrefs https://api.ahrefs.com/mcp/mcp -t http, then authenticate from /mcp.
Pricing: Lite is listed at $129/month with "API and MCP access" (Ahrefs pricing), as of October 2026. Calls use API units from your plan.
Limits: Ahrefs says using the endpoint "via custom scripts, bridges, or standalone HTTP/JSON-RPC clients is unsupported and not permitted".
07
QuerySail
QuerySail is "a hosted, read-only MCP server for Google Search Console" at https://mcp.querysail.com/gsc/mcp. Its five tools are list_properties, get_performance, compare_performance, inspect_urls (up to 20 URLs) and list_sitemaps, and it says it "cannot submit sitemaps, request indexing or change any Search Console setting".
Auth: OAuth 2.1 sign-in to QuerySail, then connect your Google account once.
Pricing: $8 per month after a 7-day free trial, as of October 2026 (QuerySail pricing).
Which servers try to request indexing, and what are the limits?
Because the API has no indexing method, a server that "requests indexing" is doing one of two other things: automating the Search Console website, or calling Google's separate Indexing API. Know which before you rely on it.
08
gsc-indexer-mcp (Mrshahidali420)
gsc-indexer-mcp reads status through the URL Inspection API (gsc_check_status, gsc_find_unindexed) and adds gsc_request_indexing and gsc_start_indexing_job. Its README is direct about how: "Google has no API for Request Indexing that a tool like this can use, so submission goes through your own browser", using a bridge extension that clicks Request Indexing in your signed-in Search Console session.
Auth: OAuth with a Google Cloud project you create; submission also needs a paired browser profile.
Limits: the README estimates "~11 slots per property" for Request Indexing on a rolling 24-hour basis; that's the project's own observation, since Google doesn't publish the number. It marks browser submission "Verified live on Windows" only, and gsc_request_indexing is "Hard-capped at five" URLs per call.
The Indexing API is the other route some servers use, including gsc-mcp above and some IndexNow servers. Google is clear about its scope: "The Indexing API can only be used to crawl pages with either JobPosting or BroadcastEvent embedded in a VideoObject" (Google). For blog posts, product pages or backlinks, that route is outside what Google supports.
Where does FreshlyCited fit next to a Search Console server?
FreshlyCited does the part a Search Console server can't: it checks whether Google has indexed a URL and, for the ones it hasn't, asks Google to look. It works on any URL, including links on sites you don't own in Search Console, and it doesn't use your Google credentials.
09
FreshlyCited
FreshlyCited checks whether Google has indexed each URL ("Seen by Google" or "Invisible"), pushes only the ones Google hasn't seen (smart push: "Already seen — not pushed"), re-checks on day 1, 3 and 7, and, for sites you verify with an IndexNow key, also sends the change through IndexNow ("Sent to Bing"). Nobody can guarantee Google indexes a page; Google decides.
Tools:check_index (index check), submit_index_job (push up to 10,000 URLs, with smart to check first), IndexNow site tools such as add_indexnow_site and verify_indexnow_site, and share_job_report for a client report link. The full MCP tool reference lists the rest.
Auth: OAuth 2.1 or an API key, to FreshlyCited, at https://freshlycited.com/mcp.
Pricing: MCP access is on Agency Pro ($199/month) and Agency Scale ($599/month). Index credits are used only if a link is indexed; if it's still not indexed after the day-7 re-check, the credit comes back automatically. A standard check costs 1 check credit. The free tier (10 free links to try, then 100 free checks a month, no card) doesn't include MCP.
Limits: it doesn't read Search Analytics, sitemaps or Search Console statuses. IndexNow reaches Bing and other participating engines, not Google (how IndexNow works).
A practical setup uses both: a Search Console server tells the agent which pages have a problem and why (see the Page indexing report statuses), the page gets fixed, and FreshlyCited checks and pushes the fixed URLs. The AGENTS.md snippet tells a coding agent to do the check-and-push step after it ships pages. More detail: FreshlyCited's API and MCP server.
Does Google offer an official Search Console MCP server?
No, as of October 17, 2026. Google's repository of official MCP servers lists remote servers such as BigQuery and Cloud Run and open-source servers including Google Analytics, but no Search Console server (google/mcp).
That's why directories show so many community options. Searching the official MCP Registry for "search-console" and "gsc" on October 17, 2026 returned more than 20 distinct Search Console servers (registry search); PulseMCP listed 33 results for "search console" (PulseMCP); Glama showed 20 on its first page (Glama). PulseMCP files some of these under "Official Providers", meaning vendors such as GSC Wizard and SEOcrawl; none is listed as a Google server.
How do you set up a self-hosted Search Console MCP server?
Setup follows the same pattern for most self-hosted servers. These steps match the service-account instructions in the mcp-server-gsc and mcp-gsc READMEs:
In Google Cloud Console, create or pick a project and enable the Search Console API.
Create a service account and download its JSON key. (For OAuth instead, create an OAuth client ID of type Desktop app.)
In Search Console, open Settings → Users and permissions and add the service account's email address to each property you want the agent to read.
Add the server to your AI client's config with the path to the key file, for example:
Restart the client and ask it to list your properties. If it returns none, the service account hasn't been added to the property yet.
Use the least access that works: a read-only scope, or "Restricted" rather than "Full" permission, if you only need reports and inspections.
How we chose
We started from the official MCP Registry, Glama and PulseMCP, then read each candidate's own repository or documentation on October 17, 2026. A server made this page if it is maintained (updated in 2026), documents its tools and auth, and represents a distinct approach: service account, OAuth, read-only, multi-engine, hosted, Indexing API, or browser-driven Request Indexing. Every tool name, auth method, price and quote comes from the project's own README, source or pricing page.
We left out servers that duplicate one already listed, servers whose repositories don't document their tools, and wrappers that resell another vendor's API through a gateway. We didn't install and benchmark each server, so we don't rate them, and GitHub stars aren't a quality score. For SEO data beyond Search Console, see our guide to SEO MCP servers for keyword, AI visibility and indexing work; for monitoring AI answers, the AI visibility trackers; and for paid indexing services, our indexing tools compared page. FreshlyCited's developer docs cover the API and MCP server in full.
A Search Console MCP server tells your agent what Google has indexed and why. When the answer is "not yet", the next step is a check and a push. Start free with 100 checks a month and see which of your pages Google has seen.
Rakesh Patel founded Space-O Technologies in 2010 and leads it as CEO, with 25 years of leadership experience; he has overseen the delivery of more than 3,000 mobile app projects. He is also the founder of TextBolt, a business email-to-SMS service, and the author of "Enterprise Mobility Strategy and Solutions" and "A Guide to Open311". Space-O built FreshlyCited to solve the "shipped but not seen" problem on its own and its clients' sites: pages and links that are live but that Google hasn't looked at yet.
Founded Space-O Technologies in 2010 · 25 years of leadership experience · author of 'Enterprise Mobility Strategy and Solutions' and 'A Guide to Open311'
Written for FreshlyCited, which helps Google and Bing find new and updated pages sooner. About FreshlyCited
Is it safe to connect an AI agent to Search Console?
It can be if you limit access. Use a server that requests the read-only webmasters.readonly scope, or give a service account Restricted rather than Full permission, and leave destructive tools such as delete_site switched off.
Do I need a Google Cloud project for a Search Console MCP server?
For most self-hosted servers, yes: you enable the Search Console API and create a service account or an OAuth client. Some servers ship their own OAuth client, and hosted servers such as QuerySail handle the Google sign-in for you.
Can I use Google's Indexing API to get blog posts indexed through MCP?
Google documents the Indexing API only for pages with JobPosting or BroadcastEvent structured data. Sending other pages to it is outside what Google supports, so use sitemaps, Request indexing in Search Console, or an indexing tool instead.
Does FreshlyCited read my Search Console data?
No. FreshlyCited checks by URL whether Google has indexed a page and pushes the ones it hasn't seen, without your Google credentials. Pair it with a Search Console MCP server if your agent also needs clicks, queries or Page indexing statuses.
