nosnippet: Block Snippets and AI Answers

How do you block Google Search snippets?

Google Search snippets are controlled with three settings: nosnippet, data-nosnippet, and max-snippet. nosnippet blocks all snippets for the page, data-nosnippet excludes only the text you specify, and max-snippet limits the length of the snippet.

To keep only certain paragraphs, not the whole page, from being excerpted, the right method is to add the data-nosnippet attribute to those parts. If you want to reduce only featured snippets, the guidance is to set max-snippet to a short length.

To remove the page itself from search results, use noindex. Unlike snippet controls, noindex removes search visibility altogether, so use each according to its purpose.

Can I block use only in Google AI Overviews and AI Mode?

AI features in Google Search are controlled with the same methods as regular search. Google says control over AI Overviews and AI Mode is done through robots.txt rules for Googlebot and through nosnippet, data-nosnippet, max-snippet, and noindex.

Google-Extended is often confused with these. Google-Extended is a control token for Gemini model training and grounding, and it does not affect inclusion or ranking in Google Search. It has no separate HTTP user agent and is used only as a robots.txt token.

In other words, blocking Google-Extended is not a way to control AI Overviews in Google Search. Also keep in mind that if you use snippet rules to reduce use in AI features, snippets in regular search results may be affected too.

How does this apply to ChatGPT and Claude?

Set separate robots.txt rules for each bot name the companies have published. Training bots and search bots are separate, so you can set things up so your site appears in search but is not used for training.

For OpenAI, GPTBot is the bot that collects content that may be used for model training, and OAI-SearchBot is the bot for showing sites in ChatGPT search results. The two settings are independent, so you can put 'User-agent: GPTBot' and 'Disallow: /' in robots.txt and still allow OAI-SearchBot.

Anthropic distinguishes ClaudeBot for training, Claude-SearchBot for search, and Claude-User for user requests, and says its bots follow robots.txt. Bot names and behavior are as of writing (2026) and may change, so check each company's official documentation.

Does blocking with robots.txt block everything?

No, it cannot be considered a complete block. Following robots.txt rules is voluntary for crawlers, and the rules may not apply to agents that fetch pages at a user's request.

OpenAI explains that robots.txt rules may not apply to ChatGPT-User because its actions are initiated by users, and Perplexity says Perplexity-User generally ignores robots.txt rules.

Google also explains that robots.txt is not a way to hide a page from Google. If the content must not be public, the right approach is to use noindex or protect it with a password.

What should you weigh before blocking?

Blocking is a decision that trades away visibility, so first identify whether what you want to stop is snippet excerpts, citation in AI answers, or model training.

Answer engine optimization (AEO) and generative engine optimization (GEO) aim to get content used in answers, which is the opposite direction from blocking. Visits earned through search engine optimization (SEO) may also be affected, so the common approach is to apply restrictions only to paid content or certain sensitive areas and leave the rest open.