Google-Extended Doesn't Control AI Overviews
Blocking Google-Extended does not keep a page out of Google’s AI Overviews or AI Mode; Search’s own snippet and indexing controls do. Google’s crawler documentation, last updated 15 October 2026, says Google-Extended does not affect a site’s inclusion in Google Search, and its AI features guide names four controls for Search: nosnippet, data-nosnippet, max-snippet and noindex.
Mercy Tahitoe, Studio Head, White Wood, Jakarta.
1. What Google-Extended actually switches off
Google-Extended is a permission token for Gemini, not a crawler you can block from Search. Google’s documentation says it has no user agent of its own: the crawling is done by Google’s usual crawlers, and the token in robots.txt only decides whether that content may be used to train future Gemini models and to ground answers in the Gemini app and in Grounding with Google Search on Vertex AI. The same page says it is not a ranking signal and does not affect inclusion in Search.
So a site that disallows Google-Extended can still be crawled, indexed, ranked, and shown inside an AI Overview.
2. The controls Search actually follows
AI Overviews and AI Mode are part of Search, so they follow the same controls as a normal search result. Google’s AI features guide says there are no extra requirements to appear in them. A page needs to be indexed and eligible for a snippet to be shown as a supporting link. To limit what Google shows from a page, the guide points to four controls, and the robots meta tag documentation spells out what each one does:
| What you want | Control | What else it changes |
|---|---|---|
| Keep a page out of Search entirely, AI features included | noindex | The page disappears from normal results too |
| Show the page but stop its text feeding AI Overviews and AI Mode | nosnippet | No text snippet in normal results either |
| Cap how much text can be used | max-snippet:[number] | The same cap applies to normal snippets; max-snippet:0 equals nosnippet |
| Hide one passage | data-nosnippet on a span, div or section | That passage never shows as a snippet anywhere |
| Stop Gemini training and Gemini grounding | Disallow Google-Extended in robots.txt | Nothing in Search |
None of the controls Google documents removes you from AI Overviews while keeping your normal snippet. Each one trims both.
Two details catch people out. Google reads these tags only if its crawler can open the page, so blocking the page in robots.txt also hides your noindex from Google. And when rules conflict, Google applies the more restrictive one.
3. Other assistants have their own switches
ChatGPT and Perplexity separate search from training, each with its own crawler name. OpenAI’s documentation describes OAI-SearchBot as the crawler for ChatGPT search and GPTBot as the crawler for training, set independently. A site that blocks OAI-SearchBot will not be shown in ChatGPT search answers, though it can still appear as a navigational link. OpenAI says robots.txt changes take about 24 hours to apply.
Perplexity’s documentation says PerplexityBot surfaces and links sites in its search results and is not used to train models. Perplexity-User, which fetches a page when a user asks for it, generally ignores robots.txt because the request comes from a person.
4. What we chose for whitewood.ai
We allow every one of these crawlers, training ones included, because being found matters more to a small studio than holding back our text. The whitewood.ai robots.txt names OAI-SearchBot, GPTBot, PerplexityBot, Perplexity-User, Google-Extended and the others one by one, and allows all of them. In SOLEDAD, White Wood’s AI-monitoring engine, our home page was cited 21 times by Google AI Overview across 3 of our 10 tracked prompts between 18 September and 18 October 2026.
Your answer might differ. A publisher whose income depends on people reading the full article may decide nosnippet on some pages is worth the lost snippet. Make that call page by page, not for the whole domain.
If you change a setting, check the page in Search Console’s URL Inspection tool. Google’s guide warns that a recrawl can take from several days to several months.
What this cannot tell you
These controls decide what an engine may use, not whether it will use you. Allowing every crawler did not put whitewood.ai into most answers: our home page citations came from one engine on 3 prompts. We have not tested nosnippet on our own pages, so we cannot show you what removing a page from AI Overviews does to its clicks. Controls also change; check the date on each source below before you act.
Declared interest: we sell this work at White Wood and track the results in SOLEDAD, the AI-monitoring engine we built.
FAQ
If I block Google-Extended, will my rankings drop?
Google says Google-Extended is not used as a ranking signal and does not affect inclusion in Search.
Can I keep my page in Google but out of ChatGPT search?
Yes. Disallow OAI-SearchBot in robots.txt. OpenAI says the page can still appear as a navigational link.
Sources
- Google Search Central, Google’s common crawlers (Google-Extended), last updated 15 October 2026: https://developers.google.com/crawling/docs/crawlers-fetchers/google-common-crawlers
- Google Search Central, AI features and your website: https://developers.google.com/search/docs/appearance/ai-features
- Google Search Central, robots meta tag specifications, last updated 24 March 2026: https://developers.google.com/search/docs/crawling-indexing/robots-meta-tag
- OpenAI, crawler documentation: https://developers.openai.com/api/docs/bots
- Perplexity, Perplexity crawlers: https://docs.perplexity.ai/guides/bots
- whitewood.ai robots.txt, read 26 October 2026: https://whitewood.ai/robots.txt
- SOLEDAD, White Wood’s AI-monitoring engine: whitewood.ai baseline, 10 unbranded prompts on six engines, 18 September to 18 October 2026
