Google · Model training
Google-Extended: Google's Gemini opt-out token, and its limits
Google-Extended is a robots.txt product token, not a crawler: it tells Google whether content it has crawled may train future Gemini models and ground answers in Gemini Apps and Vertex AI. No request ever carries it; the fetching is done by Googlebot and Google's other existing user agents. It has no effect on Google Search inclusion or ranking, and it does not cover AI Overviews or AI Mode, which are Search features.
- robots.txt token
Google-Extended- User agent
- None: a robots.txt token only, it never visits on its own
- Operator
- Type
- Model training
- What it feeds
- Training of future Gemini models (Gemini Apps, Vertex AI API for Gemini) and grounding in Gemini Apps and Grounding with Google Search on Vertex AI.
- Follows robots.txt
- Yes. It exists only in robots.txt: Google says the Google-Extended token "is used in a control capacity" and has no HTTP user agent of its own.
- How to verify it
- Not applicable: there is no separate crawler. Fetches come from existing Google user agents such as Googlebot, verified by reverse DNS or Google's published IP ranges. (list)
Should you block Google-Extended?
Disallowing Google-Extended opts your content out of Gemini training and of grounding in Gemini Apps and Vertex AI, with no effect on Google Search inclusion or ranking. It does not stop Googlebot crawling, and it does not keep you out of AI Overviews or AI Mode. For most businesses it is the low-cost choice if you object to Gemini training.
User-agent: Google-Extended
Disallow: /
User-agent: Google-Extended
Allow: /
Add these lines to /robots.txt at your site root. A group naming Google-Extended replaces your User-agent: * rules for it, so repeat any paths you also want it to skip.
Live check
Is Google-Extended allowed on your site?
Google-Extended and Googlebot
Google-Extended only makes sense next to the crawler that does the actual work:
| Token | Sends requests? | What it controls |
|---|---|---|
| Googlebot | Yes | Crawling for Google Search (including AI Overviews and AI Mode), Discover, Images, Video mode, News |
| Google-Extended | No | Whether crawled content may train future Gemini models and ground answers in Gemini Apps and Grounding with Google Search on Vertex AI |
Google's docs put it directly: "Google-Extended doesn't have a separate HTTP request user agent string. Crawling is done with existing Google user agent strings." So a Google-Extended rule is a usage instruction attached to content Google already fetches.
What disallowing Google-Extended covers
Per Google, the token manages whether content may be used for:
- training future generations of Gemini models that power Gemini Apps and the Vertex AI API for Gemini
- grounding, which Google defines as providing content from the Google Search index to the model at prompt time, in Gemini Apps and Grounding with Google Search on Vertex AI
What it does not cover
- Google Search. "Google-Extended does not impact a site's inclusion in Google Search nor is it used as a ranking signal in Google Search."
- Crawling. Googlebot keeps fetching your pages. If you need to reduce Google's crawl load, that is a Googlebot matter.
- AI Overviews and AI Mode. These are Google Search features, so they depend on Googlebot crawling, and Google's description of Google-Extended does not mention them. In robots.txt terms they follow your Googlebot rules, and blocking Googlebot takes you out of Search as a whole.
That last point is the most common misunderstanding: adding Google-Extended to robots.txt is a Gemini decision, not a Search AI decision.
How to set it
Add a group at the root of each host:
User-agent: Google-Extended
Disallow: /
Or opt out only one section, using the same path syntax Google shows in its example:
User-agent: Google-Extended
Allow: /archive/1Q84
Disallow: /archive/
A Google-Extended group does not change your Googlebot group, so you don't need to repeat Search rules inside it.
Verifying
There is nothing to verify for the token itself. The requests are Googlebot's (or another Google crawler's), and you verify them the usual way: reverse DNS to a googlebot.com hostname, or a match against common-crawlers.json. You can check whether your robots.txt names Google-Extended correctly with the robots.txt checker.
Questions
- Why don't I see Google-Extended in my server logs?
- Because it never sends requests. Google says Google-Extended has no separate HTTP user agent string and crawling is done with existing Google user agents. You'll see Googlebot and others; Google-Extended only changes how Google may use what they fetched.
- Does blocking Google-Extended hurt my Google rankings?
- No. Google states that Google-Extended does not impact a site's inclusion in Google Search and is not used as a ranking signal.
- Does Google-Extended keep my site out of AI Overviews?
- No. AI Overviews and AI Mode are Google Search features, so they follow Googlebot's rules, not Google-Extended's. Google's description of the token covers Gemini training and Gemini or Vertex AI grounding only.
- Can I opt out of Gemini training for only part of my site?
- Yes. Google-Extended takes normal Allow and Disallow paths, so you can disallow a directory such as /archive/ and leave the rest open.
- Do I need a rule on every subdomain?
- Yes. robots.txt is read per host, so each subdomain needs its own /robots.txt containing the Google-Extended group.
Related crawlers
Sources
Checked against these pages on 2026-10-18. Operators change their crawlers; check the source before relying on a detail.