24
Crawlers documented
11
Operators
7
Fetch on a user's request
Search index
Builds the index an AI answer searches. Block it and you drop out of that product's results.
User-triggered fetch
Visits a page because a person asked a question that needs it. This is the visit that can end in a citation.
Model training
Collects pages for training future models. Slower to pay off, and separate from search and citations.
Named in our robots.txt, undocumented
Our own robots.txt also names Claude-Web, anthropic-ai, Bytespider, YouBot, cohere-ai, GrokBot and xAI-Grok. We could not find current documentation for these tokens from the companies they are attributed to, so they have no page here: a page would have to guess what they do.