Does a noindex tag keep your pages out of AI search results? For Google's AI Overviews and AI Mode, yes — Google's own documentation says a page must be indexed and snippet-eligible to appear as a supporting link, so noindex removes it completely. For ChatGPT, Claude and Perplexity, noindex is mostly irrelevant: those systems run their own crawlers, and robots.txt is the documented control. And since late 2026, Google offers a third option — a Search Console toggle that opts you out of AI Overviews without touching your classic rankings. Here is the honest breakdown, engine by engine.
Google: noindex means no AI Overviews, full stop
Google's AI features and your website documentation states the rule in one sentence:
To be eligible to be shown as a supporting link in AI Overviews or AI Mode, a page must be indexed and eligible to be shown in Google Search with a snippet, fulfilling the Search technical requirements.
Read it as three gates: Googlebot must be able to crawl the page, the page must be indexed, and it must be eligible to show a snippet. A noindex tag fails gate two. The page leaves Google's index, and with it any chance of appearing in AI Overviews or AI Mode. There is no separate AI index to worry about and no special markup that overrides it.
The snippet gate matters too, and it is subtler. Google's robots meta tag documentation confirms that nosnippet prevents a page's content from being used as direct input for AI Overviews and AI Mode, and max-snippet caps how much of it can be used. The data-nosnippet attribute does the same for individual elements. So the control ladder for Google looks like this:
- noindex — page leaves the index, gone from Search and AI features.
- nosnippet — page stays indexed and ranking, but its text can't feed AI Overviews or AI Mode directly.
- data-nosnippet — surgical: keep the page fully visible, but wall off specific passages (prices, quotes, data tables) from AI reuse.
Generate and audit these tags with our free meta tags generator — it is the fastest way to check what your pages actually declare.
The classic noindex gotcha still applies
You must be able to crawl a page to read its noindex directive. If you block the page in robots.txt, Googlebot never fetches it, never sees the tag, and the URL can linger in results — sometimes for months — with only the URL showing. The fix is counterintuitive but well documented: allow the crawl, serve the noindex, let Google process it, and confirm the status in Search Console's URL Inspection tool. Never use robots.txt and noindex on the same URL and expect the tag to work.
Google's new option: opt out of AI features without noindex
Until recently, publishers faced an all-or-nothing choice: noindex removed you from AI Overviews but also from the classic search results that still drive most traffic. In late 2026 Google separated the two. Search Console now has a Settings → Search generative AI control covering AI Overviews, AI Mode and generative AI features in Discover. How it works, per Google's documentation as reported this week: inclusion is the default. When enabled, your content can appear as links in AI features and help ground generated responses, which can generate impressions and traffic. Excluding a property reverses that — your content is no longer shown in those surfaces as links or used to ground generated responses — while your classic search presence is untouched.
This is the control news publishers and specialist sites have asked for since AI Overviews launched: decline the AI summary, keep the blue links. If your business model depends on the visit rather than the citation, this toggle is almost certainly a better fit than noindex. Two practical notes: the setting inherits from parent properties to child properties in Search Console, so check what your domain-level property passes down; and Google says it expects to extend the covered surfaces over time, so revisit the setting periodically rather than treating it as set-and-forget.
For the full walkthrough of opting out while staying indexed, see our guide to opting out of Google AI Overviews.
What about Google-Extended?
A quick clarification because this token causes constant confusion: Google-Extended is not a search control. Google's crawler documentation says it does not affect a site's inclusion in Google Search and is not a ranking signal. It controls whether your content can be used for AI training and grounding in Google's other systems (Gemini apps, Vertex AI). Blocking it does nothing to your AI Overview eligibility — that is governed by indexing and snippet rules. We covered the token in detail in our Google-Extended guide.
ChatGPT, Claude, Perplexity: noindex is the wrong tool
Here's where the honest answer gets uncomfortable for anyone who hoped one tag would rule everything. OpenAI, Anthropic and Perplexity operate their own crawlers, and none of them documents the noindex meta tag as a control:
- OpenAI documents robots.txt tokens: disallowing
GPTBotkeeps content out of foundation-model training; opting out ofOAI-SearchBotremoves pages from ChatGPT search answers. Its crawler documentation does not mentionnoindex. - Anthropic documents three robots.txt tokens —
ClaudeBot,Claude-SearchBot,Claude-User— all of which honor robots.txt. Nonoindexhandling is documented. - Perplexity documents
PerplexityBotvia robots.txt. Same story: no publicnoindexcommitment.
Could these systems respect noindex voluntarily behind the scenes? Possibly — any serious crawler parses meta tags — but "possibly" is not a control you can audit, verify, or rely on. If you want content out of ChatGPT's answers, block OAI-SearchBot in robots.txt. If you want it out of training, block GPTBot. The meta tag is not the mechanism, documented or otherwise.
Bing and Copilot: index-based, like Google
Microsoft's Copilot answers are grounded in the Bing index, which makes Bing behave like Google for this question: a page excluded from Bing's index (via noindex, honored by Bingbot) has no route into Copilot's citations. The practical wrinkle is that you must manage it in the Bing ecosystem — Bing Webmaster Tools, Bingbot crawl — not just Google's. Sites that carefully noindex pages for Google but never check Bing's index of the same URLs are half-done.
The decision guide: which control for which goal?
- "I don't want this page in Google's AI answers, but I want rankings." — Search Console's Search generative AI toggle. No code changes, classic search untouched.
- "I don't want this page indexed anywhere." —
noindexmeta tag (orX-Robots-Tagfor non-HTML files). Remember: allow the crawl so the tag gets read. - "I want the page indexed, but AI shouldn't quote its text." —
nosnippetfor the whole page, ordata-nosnippetfor the specific passages. - "I don't want OpenAI/Anthropic training on my content." — robots.txt tokens (
GPTBot,ClaudeBot,Google-Extended). Meta tags don't do this. - "I don't want to be cited in ChatGPT/Claude answers." — robots.txt tokens for the answering bots (
OAI-SearchBot,Claude-SearchBot,Claude-User). Again, not a meta-tag problem.
Notice the pattern: Google's world is controlled with meta tags and Search Console; everyone else's world is controlled with robots.txt. The sites that get this wrong are the ones that apply one company's mechanism to another company's crawler.
Audit your setup this week: the 20-minute checklist
Theory is nice; here is the practical pass to run on your own site:
- List every noindex on the site. In Search Console, check the Pages report for "Excluded by 'noindex' tag" and confirm each URL is there deliberately. Stale noindex tags from old staging migrations are one of the most common silent traffic killers.
- Check snippet directives. Search your templates for
nosnippetandmax-snippet— including any SEO plugin's global settings. A site-widemax-snippet:0set years ago for a different reason will quietly starve your AI Overview eligibility today. - Decide on the Search Console toggle. Visit Settings → Search generative AI for each property. If you're a publisher whose business is the visit, consider excluding AI surfaces while keeping classic search — but make it a deliberate decision, documented somewhere, not a default you never questioned.
- Verify the non-Google side separately. Open your robots.txt and confirm the training tokens (
GPTBot,ClaudeBot,Google-Extended) and answering tokens (OAI-SearchBot,Claude-SearchBot,PerplexityBot) each have the stanza you intend. Remember: noindex does nothing here — this file is the entire control plane. - Spot-check Bing. If you changed noindex or robots.txt recently, confirm Bing Webmaster Tools shows the same state as Google. Copilot citations follow Bing's index, and the two search engines don't sync your directives for you.
Run this quarterly. Directives drift — plugins update, templates change, someone "temporarily" noindexes a page and forgets — and each drifted directive is either lost traffic or lost protection.
FAQ
Does noindex remove my page from Google AI Overviews?
Yes. Google requires pages to be indexed and snippet-eligible to appear as supporting links in AI Overviews or AI Mode. A noindexed page fails the indexing requirement and cannot appear.
Does noindex stop ChatGPT from citing my content?
There is no documented mechanism for that. OpenAI's published controls are robots.txt tokens: GPTBot for training, OAI-SearchBot for search answers. If you want out of ChatGPT citations, disallow OAI-SearchBot in robots.txt.
Can I opt out of AI Overviews without hurting my Google rankings?
Yes — that is exactly what the Search Console "Search generative AI" setting (Settings → Search generative AI) is for. It excludes your property's content from AI Overviews, AI Mode and generative AI features while leaving classic search inclusion and rankings untouched.
What's the difference between noindex and nosnippet for AI search?
Noindex removes the page from the search index entirely — no rankings, no AI citations. Nosnippet keeps the page indexed and ranking but prevents its text from being used as direct input for AI Overviews and AI Mode. Use data-nosnippet to protect individual passages while leaving the rest of the page fully usable.
Why is my noindexed page still showing in search results?
Almost always because robots.txt blocks Googlebot from crawling the URL. Google can't read a noindex tag it never fetches. Allow the crawl, serve the tag, and check the page's status in Search Console's URL Inspection tool.
Does Google-Extended affect AI Overviews?
No. Google-Extended is a training opt-out for Google's other AI systems (Gemini apps, Vertex AI). It has no effect on Search inclusion, rankings, or AI Overview eligibility, which are governed by indexing and snippet rules. Plus Google's new Search Console opt-out that preserves rankings.
