Google-Extended
Gemini model training and grounding
What Google-Extended does
Google-Extended is not a crawler and has no user agent of its own. Nothing visits your site calling itself Google-Extended. It is a token you name in robots.txt to withhold permission for content Google has already crawled to be used for Gemini training and grounding.
That distinction is the reason it is so widely misunderstood. People reason from the shape of the name, assume it must be a crawler like the others, and conclude that disallowing it removes them from Google's AI features. It does not. Google states that it does not affect Search ranking, and it does not remove you from AI Overviews, which are built from the ordinary index.
In our survey, 32 of 148 readable files disallowed it, and news publishers were the group doing it — the ones with the clearest commercial reason to withhold training permission while staying in Search.
If you want the training opt-out, this is the correct token and it costs you nothing in visibility. If you want out of AI Overviews, this is not the mechanism and there is not one.
Google-Extended is not Googlebot
Google-Extended is not a crawler. It is a permission token attached to crawling Googlebot has already done.
What Googlebot does insteadHow to block Google-Extended
Name the token in robots.txt. There is no crawler to turn away, so this withholds a permission rather than refusing a request.
robots.txt
User-agent: Google-Extended
Disallow: /What that costs you: Not your ranking, and not AI Overviews. This token is widely believed to control AI Overviews and does not: those come from the ordinary index.
To allow it instead, write nothing. A crawler with no group of its own follows your User-agent: * rules, and giving it a named group with Allow: / detaches it from those rules entirely — including any path you disallow there.
Find it in your logs
Nothing to find. No request ever arrives identifying itself as Google-Extended.
Check that it is really Google-Extended
There is nothing to verify. No request ever arrives identifying itself as Google-Extended: the token is read out of your robots.txt by Google’s ordinary crawler, on their behalf, and applied to crawling that has already happened. Anything in your logs claiming to be Google-Extended is not what it says it is.
Should you block Google-Extended?
A clean choice with no visibility cost. Google states it is not a ranking signal and it does not remove you from AI Overviews, so if you want the Gemini training opt-out this is the correct token and it is free.
What it will not do is the thing most people disallowing it are hoping for. It does not keep you out of AI Overviews. Seven of the ten sites we found blocking it were news publishers, who generally do want the training opt-out and do want to stay in Search, and for them it is exactly right.
Questions
- Does blocking Google-Extended hurt my rankings?
- No. Google states it is not a ranking signal. It withholds permission for training and grounding, and leaves crawling and ranking alone.
- Why does it have no IP range list?
- Because nothing crawls under that name. There is no traffic to verify. The token exists to be matched in robots.txt, which Googlebot reads on Google's behalf.
The user agent and IP figures on this page were read from Google’s own URLs on 14 September 2026. Blocking figures come from our survey of 170 well-known sites across 10 categories on 14 September 2026, 148 of which returned a readable robots.txt.
Is Google-Extended allowed on your site right now?
The checker reads your live robots.txt and reports every token in this directory, with the exact directive that decided each one.