In short. Applebot-Extended lets publishers opt web content out of Apple foundation model training while preserving separate Applebot crawl permissions.

Definition: As of 17 October 2026, Applebot-Extended is Apple's robots.txt token for opting website content out of training its general purpose foundation models. It governs use of data collected by Applebot, Apple's web crawler.

Key facts

Apple's About Applebot documentation states that Applebot-Extended performs no crawling and that disallowed pages can remain in search results.
Apple's foundation model overview describes training on licensed data and publicly available data collected by Applebot, with a publisher opt-out for web content.
The IETF's Robots Exclusion Protocol specifies a lowercase /robots.txt file at the service's top-level path, encoded as UTF-8 and served as text/plain.

What use does the token control?

Apple's data usage section covers training models powering Apple Intelligence, Services, and Developer Tools. The same documentation states that Applebot-Extended rules are excluded from Search ranking.

Use in AI answers has a separate control: nosnippet. Apple documents this directive for excluding content from the additional context supplied to models when generating output in Apple products and services. A training opt-out and an answer-content restriction therefore require separate settings.

How is the robots.txt rule written?

Apple's documented example excludes the /private/ path from training use:

User-agent: Applebot-Extended
Disallow: /private/

For an exclusion covering all paths, the rule becomes:

User-agent: Applebot-Extended
Disallow: /

The protocol specification defines path matching from the beginning of the path. The slash in the second example consequently matches every path. These examples express alternative scopes.

A robots.txt group consists of user-agent lines followed by rules. The User-agent line identifies the token to which the following Disallow applies. A new user-agent line after those rules begins another group. Keeping the Applebot-Extended group separate makes its intended scope explicit.

How can the published setting be checked?

The IETF's access-method requirements provide concrete checks: retrieve /robots.txt from the service root, confirm the filename's lowercase spelling, and inspect its encoding and media type. Then inspect the group containing Applebot-Extended and its path rule.

The same standard requires matching groups for the same token to be combined. Its path rules use the most specific match, so reviewing every applicable Allow and Disallow matters when checking an exclusion. The protocol recommends case-sensitive path matching; capitalization in a path can therefore change which rule applies.

FAQ

Does blocking Applebot-Extended stop Apple search crawling?

Apple's documentation separates Applebot crawling from Applebot-Extended data-use rules. Your Applebot crawl permissions remain a separate setting.

Can I exclude only part of a website?

Yes. Apple's example uses Disallow: /private/ to define a path-specific exclusion. You can compare that scope with the all-paths rule shown above.

Sources

Check how a page reads to answer engines. Open the GEO score analyzer