Amazon · Training
Amazonbot
Of Amazon's three tokens this is the only one carrying no training carve-out, and the omission is the point: Amzn-SearchBot and Amzn-User both state they do not crawl content for generative AI model training, while this one is described as improving Amazon's products and services and "may be used to train Amazon AI models" (note the hedge). It sweeps on Amazon's own schedule rather than firing on a customer action. Verification is weaker than the range count suggests: the 524 published addresses are bare IPv4 with no CIDR suffix under an `ipv4Prefix` key, wrapped in an HTML page rather than served as JSON, and there is no IPv6 and no documented reverse DNS, so an Amazonbot request arriving over IPv6 cannot be checked at all.
- Operated by
- Amazon
- Purpose
- Training
- robots.txt token
- Amazonbot
- Verification
- IP verified
Collects pages into a corpus used to train models.
How to verify Amazonbot
Amazon publishes the IP ranges Amazonbot crawls from, and we fetch that list on a schedule. A request claiming to be Amazonbot can therefore be checked against 1,292 published ranges: one that does not match is not this crawler.
Published ranges last confirmed by us on .
User agent
Amazon publishes this user agent for Amazonbot. Match on the Amazonbot product token rather than the whole string: vendors revise the surrounding version and URL fragments without notice.
Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; Amazonbot/0.1) Chrome/W.X.Y.Z Safari/537.36robots.txt for Amazonbot
Yes, and Amazon is unusually precise about it: RFC 9309 is cited by name, the user-agent and allow/disallow directives are honoured, `rel=nofollow` and the `noarchive`, `noindex` and `none` meta tags are supported, and `crawl-delay` explicitly is not. Three operational traps. Robots.txt is fetched per host, so rules at your apex do not cascade to `site.example.com`. Amazon may serve itself a cached copy from the last 30 days and fails open (when a file can't be fetched it behaves as if it does not exist), with changes taking around 24 hours to reflect. And Amazon glosses `noarchive` inline as "do not use the page for model training," which is not what that tag means to any other search engine, so an operator who set it years ago for cache reasons has silently opted out of Amazon training.
Block
User-agent: Amazonbot
Disallow: /Allow
User-agent: Amazonbot
Allow: /robots.txt is a request, not an enforcement mechanism. It is honoured by convention, and a crawler that ignores it is stopped at your edge, not in a text file.
What blocking Amazonbot costs you
Amazon is the only vendor in this directory attaching cash to an allow: the same page advertises Amazon Content Partners, where allowing this token in robots.txt "may" make you eligible for a +1% affiliate commission boost on eligible sales, free hosting credits and AI traffic management tools. For an affiliate-revenue site that is a quantifiable line item rather than a hypothetical, though the eligibility terms sit behind a separate programme page. Everything else you forgo is invisible: exclusion from Amazon's model training inputs produces no referral traffic and no analytics signal in either direction. Note also that the three tokens' settings are independent, so blocking this one leaves the other two untouched.