Google-Extended: definition and what it controls

99Visibility Research · Updated October 21, 2026 · 4 min read

Short answer

Google-Extended is a robots.txt token that tells Google whether your content may be used to train Gemini models and to ground Gemini Apps and Vertex AI answers. Google states it does not affect your inclusion in Google Search and is not a ranking signal. It also does not control AI Overviews or AI Mode, which follow your Googlebot rules.

Key takeaways

Google-Extended governs Gemini training and grounding for Gemini Apps and Vertex AI, not Google Search.
Google says blocking Google-Extended does not change Search inclusion and is not used as a ranking signal.
There is no robots.txt token for AI Overviews or AI Mode; Googlebot rules and Search Console controls govern those.
No published study measures whether blocking Google-Extended changes what Gemini says about a brand.

Definition

Google-Extended is a product token you can name in robots.txt. It sets whether content Google has crawled may be used for its generative AI products outside Search. Google lists it in its common crawlers documentation, last checked in our evidence as of the July 2026 revision.

As of October 2026, Google describes two uses it governs:

Training future Gemini models.
Grounding answers in Gemini Apps and Vertex AI, meaning pulling your pages in to support a live answer.

It is a permission flag, not a visibility lever. It tells Google what it may do with your content. It does not make Google show your content anywhere.

How it works

You write rules for Google-Extended in robots.txt the same way you would for any user agent. A site that wants to opt out of Gemini training and grounding entirely adds:

User-agent: Google-Extended
Disallow: /

Your Googlebot rules stay separate. Googlebot still decides whether pages can be crawled and indexed for Search. Google-Extended only decides the downstream Gemini uses. For the side-by-side detail, see Google-Extended vs Googlebot.

Three things this token does not touch:

AI Overviews and AI Mode. Google says AI is built into Search, so Googlebot directives are the control. Eligibility requires a page to be indexed and snippet-eligible, with no extra requirements.
Search ranking. Google states the token is not used as a ranking signal.
Content already used. A model trained before you added the rule has already frozen what it knows at its cutoff.

Why it matters for AI visibility

Google runs several generative surfaces, and each has a different control. Mixing them up is the most common way site owners block the wrong thing.

To keep content out of Gemini training and Gemini Apps grounding, use Google-Extended.
To limit how much of a page appears in AI Overviews or AI Mode, use the snippet controls Google documents: nosnippet, max-snippet, data-nosnippet and noindex.
To exclude a property from Google's generative features in Search, use the Search Console control described in Google's AI features opt-out guide. Inclusion there is the default, so check it was not switched off.

If your goal is to refuse training while staying visible in AI answers, the block training but stay in AI search playbook walks through the per-bot setup across Google, OpenAI and others. For Gemini specifically, Gemini crawler controls covers the token alongside firewall settings.

Common misconceptions

Blocking Google-Extended removes you from AI Overviews

It doesn't. AI Overviews draw on the Search index, which Googlebot governs. The full answer is in does blocking Google-Extended remove my site from AI Overviews.

Blocking it hurts your Google rankings

Google says it does not affect Search inclusion and is not a ranking signal. Google's statements are assertions rather than independent measurement, but no evidence contradicts them. See whether Google penalizes blocking AI crawlers.

Allowing it makes Gemini recommend you

No source shows that. Training effects on what a model recalls about a brand are unmeasured by anyone, including Google. Allowing the token grants permission; it promises nothing about answers.

Frequently asked questions

Does blocking Google-Extended change what Gemini says about my brand?

Not in any way anyone has measured. Blocking stops future training and Gemini Apps grounding use, but models already trained keep what they learned up to their cutoff. Whether training exclusion changes brand recall has not been tested by any published study. The longer answer is in [this question on Gemini brand answers](/questions/google-extended-gemini-brand-answers).

Is Google-Extended a separate crawler I can see in my logs?

Treat it as a robots.txt token that governs how Google may use content, not as a bot you need to allow through your firewall separately. Your Googlebot access rules still decide whether Google can fetch pages at all. If Googlebot is blocked, the Google-Extended setting matters little because Google cannot crawl the content in the first place.

Can I block Google-Extended on only part of my site?

Yes. Robots.txt rules work per path, so you can disallow Google-Extended for specific directories, such as a research library, and allow the rest. Keep the rules simple and test them. Accidental blanket disallows on the wrong user agent are a common source of silent visibility problems.

Does Google-Extended control the Gemini API's Grounding with Google Search?

Our evidence treats the Gemini API's developer-invoked Grounding with Google Search as a separate Google retrieval path from Gemini Apps. Google documents Google-Extended for Gemini Apps and Vertex AI. If this path matters to you, read Google's current crawler documentation directly rather than assuming one token covers every Gemini product.

Sources

This page is AI-assisted. It was written from a verified evidence base and checked against our editorial policy. Engine behaviour changes often; facts are as of October 21, 2026.