How to remove a page from AI search, engine by engine
Remove a page from AI search: the control, the stated delay and the check for Google AI Overviews and AI Mode, Bing and Copilot, ChatGPT search and Perplexity.
By Crawlito · 9 min read
In short
Google's Search generative AI control in Search Console excludes a property from AI Overviews, AI Mode and Discover's generative AI features, and Google says content is generally excluded within a few days.
Google's Removals tool hides a URL from Search for about six months, usually takes up to a day to process, and its help page does not mention AI features, so it is a stopgap, not the fix.
Bing's new ai-grounding:disallow page preference keeps a page out of AI answers while leaving it in Bing results, but Microsoft only starts honouring it as it crawls from 1 February 2027.
OpenAI says sites that disallow OAI-SearchBot are not shown in ChatGPT search answers, may still appear as navigational links, and that robots.txt changes take about 24 hours.
Perplexity documents only robots.txt for PerplexityBot, says changes can take up to 24 hours, and says Perplexity-User generally ignores robots.txt.
To remove a page from AI search, you change a signal each engine reads and then wait for that engine to fetch it again. Google has two dashboard controls: the Search generative AI control for whole properties and the Removals tool for urgent single URLs. Bing is moving Copilot to a new page-level grounding preference. ChatGPT search and Perplexity document only robots.txt rules, plus a noindex tag in OpenAI's case. The one change that works whatever an engine documents is to stop serving the content: a 404, a 410 or a login.
Pick the outcome before you touch anything, because the controls conflict. A robots.txt Disallow stops crawlers from seeing a noindex tag on the same URL, so stacking both can leave a page half removed.
Outcome you need
Change at the source
Engine-side controls that help
Gone from the web and every engine
Return 404 or 410, or put it behind a login
Google Removals, Bing Content Removal
Out of all search, still readable by link
noindex, page left crawlable
Google Removals for speed
In Google Search, out of Google AI features
nosnippet on the page, or exclude the property
Search generative AI control
In Bing results, out of Copilot answers
Bing grounding preference on the page
Temporary grounding preference in Bing Webmaster Tools, for sites already using noarchive or nocache
Out of ChatGPT search or Perplexity only
robots.txt group for OAI-SearchBot or PerplexityBot
None documented
If the page is sensitive, change the source first. Every engine-side tool hides results; none stops someone opening the URL.
Google: AI Overviews, AI Mode and Discover
Google's AI features page says a page must be indexed and eligible to be shown with a snippet to appear as a supporting link, so noindex takes a page out of AI Overviews and AI Mode as well as Search. Two further controls narrow the effect to AI features or snippets.
Checklist
Exclude a whole site or section from AI features. In Search Console, open Settings > Search generative AI and choose the option to exclude your site's links and content. Google's help page for the control says it covers AI Overviews, AI Mode and generative AI features in Discover, and that it was rolled out to all websites worldwide as of 31 August 2026. A URL-prefix property for a folder, such as https://example.com/old-campaign/, is a child property and can carry its own setting for that path and its subpages. How the Search generative AI control works covers inheritance and the trade-offs.
Exclude one page from AI features but keep it in Search. Add nosnippet. Google's robots meta tag reference says it also prevents the content being used as a direct input for AI Overviews and AI Mode, and it removes your ordinary snippet too. data-nosnippet hides only the marked passage from snippets, and Google's reference does not mention AI features for it.
Remove the page from Google entirely. Serve noindex on a crawlable page, or 404/410 if it is gone, then request a recrawl in URL Inspection.
If it is urgent, file a removal. In the Removals tool, Temporarily remove URL hides the URL from Search results. Google says a request usually takes up to a day, is not guaranteed to be accepted, and lasts about six months. The prefix option covers both www and non-www but not other subdomains. Clear snippet in search removes only the description, until the page is indexed again. What to do after the six months is in Google removed a URL temporarily.
Stated delay
Google says the generative AI control generally takes a few days, and content should be excluded within one to two days after the control goes live, with some content taking longer because of caching. For noindex, Google's block indexing guide warns that, depending on the page's importance, it may take months for Googlebot to revisit it, and points to the removals documentation when you need a page gone quickly.
What the Google controls do not do
The generative AI control does not stop AI training (that is Google-Extended), does not stop similar content from other sites appearing in these features, and does not override your Merchant Center or Google Ads choices. The Removals help page does not mention AI Overviews or AI Mode at all. Since Google says a supporting link must be indexed and eligible to show with a snippet, a hidden URL should not qualify, but treat that as inference, not a documented promise.
How to confirm
Use URL Inspection to see whether indexing is allowed and to view the HTML Google crawled, so you can confirm it saw your noindex or nosnippet. Check the Removals page for the request status. Then search the queries that used to surface the page in AI Mode and look for your URL among the sources.
Microsoft: Bing and Copilot
Microsoft's grounding preferences are read by Bingbot and cover Microsoft experiences including Copilot, so you work through Bing for both. The rules are changing, so check the date before you choose a tag.
Checklist
Remove the page from Bing entirely. Serve noindex or a 404/410. An older [Bing webmaster post on removing URLs](https://blogs.bing.com/webmaster/June-2009/How-to-remove-URLs-from-our-index-(expanded-editio) lists a 404 and a noindex tag among the ways to qualify for removal. Bing Webmaster Tools also has a Block URLs tool for your verified site and a Content Removal tool. Their help pages load only in a browser and could not be read for this post, so this post does not state how long their effect lasts. Make the change at the source as well.
Keep the page in Bing results but out of AI answers, from February 2027. Microsoft's October 2026 announcement adds a page-level AI Grounding preference. Microsoft says these preferences do not change eligibility for traditional Bing search results:
ai-grounding:disallowextended is the narrower value: Microsoft says it limits grounding to Microsoft-operated experiences, Copilot included, and excludes third-party grounding services. To keep a page out of Copilot, you need disallow. Grounding has no robots.txt form; it is set page by page.
Until then, use the current tag. Microsoft's September 2023 post says content tagged noarchive is not included in Bing Chat answers (Bing Chat is now Copilot), while nocache allows only the URL, title and snippet. Microsoft's October 2026 post says it stops reading these tags as training or grounding preferences after 1 March 2027, and ignores them on a site as soon as it detects the new preferences there. Bing's training and grounding preferences covers the migration.
Stated delay
Microsoft says the new preferences take effect as it crawls from 1 February 2027. From 15 December 2026, sites already using noarchive or nocache can record a temporary grounding preference in Bing Webmaster Tools, which expires on 30 September 2027, or earlier if Microsoft recrawls the page and finds its new grounding preference. Microsoft's deadline for publishing the new preferences is 1 March 2027. Microsoft documents no fixed delay for an individual page.
How to confirm
Check the tag in the raw response, then watch the cited URLs in Bing Webmaster Tools' AI Performance report. You would expect an excluded page to drop out of it after Bing recrawls the page, but Microsoft does not document how quickly.
ChatGPT search
OpenAI documents no removal tool. You work through robots.txt and the page itself.
Checklist
Decide between a crawl block and noindex. OpenAI's crawler documentation says sites opted out of OAI-SearchBot will not be shown in ChatGPT search answers, but can still appear as navigational links. At the time of writing, OpenAI's publishers FAQ points to a noindex meta tag for keeping a page out of ChatGPT search, and notes the crawler must be allowed to fetch the page to read the tag. That FAQ changes often, so reread it before you rely on it.
For a single page, use noindex. Leave OAI-SearchBot allowed and add <meta name="robots" content="noindex">. This also removes the page from every other search engine that reads the tag, so it is the right choice only if you want it out of search generally.
For a section, use a robots.txt group and accept that bare navigational links may remain:
Leave GPTBot alone for this job. GPTBot is the training opt-out and does not decide what ChatGPT search shows.
Stated delay and limits
OpenAI says it can take about 24 hours from a robots.txt update for its systems to adjust. Why that window can be longer on your side is in how long robots.txt changes take. ChatGPT-User, which fetches pages when a person asks, is not covered: OpenAI says robots.txt rules may not apply to it.
Perplexity
Perplexity's bot documentation describes one control: robots.txt rules for PerplexityBot, the crawler that surfaces and links sites in Perplexity's search results.
Expect user-triggered fetches to continue. Perplexity says Perplexity-User generally ignores robots.txt because a user asked for the page.
Remove the content at the source if it must go. Perplexity documents no meta robots support, no snippet control and no removal contact on that page, and it does not say whether a blocked URL can still appear as a bare link.
Perplexity says it may take up to 24 hours for its systems to reflect robots.txt changes.
Summary of controls and delays
Engine
Page-level control
Section or site control
Stated delay
Google AI features
nosnippet, noindex, Removals tool
Search generative AI control
A few days for the control; Removals usually up to a day
Bing and Copilot
ai-grounding:disallow (from 1 Feb 2027), noarchive until then
Block URLs tool; no robots.txt grounding rule
Applied as Bing recrawls
ChatGPT search
noindex (OpenAI FAQ)
OAI-SearchBot Disallow
About 24 hours for robots.txt
Perplexity
None documented
PerplexityBot Disallow
Up to 24 hours for robots.txt
No operator documents a way to recall an answer already given, and none of these controls touches copies of your content on other sites.
How to confirm the change is live at the source
Every engine reads what your server sends, so check the raw response before you wait on any engine:
Read the three results together. A 410 means the content is gone at the source, and each engine drops it as it recrawls. A 200 with noindex and no Disallow for that path is a clean de-indexing setup. A 200 with noindex and a Disallow means the noindex is invisible to the blocked crawler, so remove one of them.
Frequently asked questions
Is there one request that removes a page from every AI search engine?
No. Each operator has its own control, and only Google and Bing offer dashboard tools. Returning a 404 or 410, or requiring a login, is the only change that works for all of them.
Does Google's Removals tool remove a page from AI Overviews?
Google's Removals help page does not mention AI features. For a lasting result, use noindex, nosnippet or the Search generative AI control, and treat the Removals tool as a way to hide the URL while that change is picked up.
Can I keep a page in Bing search but out of Copilot?
Yes, with ai-grounding:disallow, which Microsoft says does not affect eligibility for traditional Bing results. Microsoft honours it as it crawls from 1 February 2027; until then its older noarchive tag is the documented control.
Why does ChatGPT still show my blocked page as a link?
OpenAI says sites that disallow OAI-SearchBot can still appear as navigational links. At the time of writing, its FAQ points to a noindex meta tag instead, on a page OAI-SearchBot is allowed to fetch.
Check what your page sends now
Before you file any request, confirm the page serves the signal you think it does. Run the three curl commands above, or use the free check on the page: it checks a single page and needs no account.
Filed under Strategy and audits. Topics: · noindex, ai-search, meta-robots, robots-txt, bingbot
Bing's new AI grounding meta tag and X-Robots-Tag values replace NoArchive and NoCache. What allow, disallowextended and disallow do, and the 2027 deadlines.
How nosnippet, max-snippet, data-nosnippet and noindex control AI Overviews and AI Mode, what other AI tools document, and how to set X-Robots-Tag for PDFs.
robots.txt stops crawling, noindex stops listing. Why a Disallow can hide your noindex, when to use each, and how the same logic applies to AI crawlers.