Applebot-Extended
Applebot-Extended is a robots.txt name that tells Apple whether pages Applebot read may train its AI models. It doesn't crawl anything.
Applebot-Extended is a name you can use in your robots.txt to tell Apple whether your pages may be used to train its AI models. It isn't a separate crawler. It never visits your site. Applebot still does the reading, and Applebot-Extended only decides what Apple may do with what it read.
Apple's page about Applebot explains both names.
Why it exists
Applebot reads pages for search in Siri, Spotlight and Safari. Apple also uses what it collects to help train the models behind Apple Intelligence. Some site owners are happy with the first and not the second.
Blocking Applebot would keep you out of Apple's search features too. Applebot-Extended lets you split the two: let Apple read your pages for search, and keep them out of training.
How to use it
Add a group for it in your robots.txt and disallow it. For example, Kestrel Roasters wants to show up in Siri and Spotlight but keep its pages out of Apple's training:
User-agent: Applebot
Allow: /
User-agent: Applebot-Extended
Disallow: /Apple says pages that disallow Applebot-Extended can still appear in search results. You can also block it for part of your site only, such as your blog, and leave the rest open.
Because it never visits, you won't see Applebot-Extended in your server logs. The only place it shows up is your robots.txt.
Other names like it
Google has a similar name, Google-Extended, for its own AI training controls. The idea is the same: one bot reads, and a separate name in robots.txt says whether the reading may go further.
In AnswerOrbit
The Audit tab in Website reads your robots.txt every day and shows, for each AI bot we know, whether it's Allowed, Partly allowed or Blocked, with its purpose, such as training or search.