What are Amazonbot and Amzn-User?
Amazonbot is Amazon's web crawler. Amazon documents it as improving Amazon services, with Alexa's ability to answer questions as the named example. On tollbit.com it made 2.9K successful reads between Sep 15 and Oct 14 2026, fifth among all AI agents. Amzn-User, a separate Amazon user agent that carries the "-User" suffix the industry uses for user-triggered fetchers, appeared in single digits.
Amazonbot, as Amazon documents it
- Identity. The user-agent string carries the token
Amazonbotand a link to Amazon's developer page for the crawler. - robots.txt. Amazon states that Amazonbot honors robots.txt, and the token to use is
Amazonbot. - Verification. Amazon's developer page describes reverse-DNS verification: a genuine request resolves to a hostname in Amazon's amazonbot crawler domain. A request carrying the token that does not resolve there is not Amazon.
- Purpose. Amazon's wording is "to improve our services," with Alexa answering more questions as the example. Amazon does not publish a list of which other products draw on the crawl.
Treat the last point carefully. Amazon's shopping assistant Rufus and the 2025 Alexa+ relaunch are both products that answer questions from web content, but whether either reads Amazonbot's crawl is not publicly documented.
Amzn-User, as far as the public record goes
Amazon's public bot documentation describes Amazonbot in detail. Amzn-User's exact purpose, its robots.txt handling and its verification method are not publicly documented in the same way. Its name follows the convention set by ChatGPT-User, Claude-User and Perplexity-User: a fetcher that goes to a page because a person asked an Amazon assistant something that page can answer. Until Amazon documents it, treat it as retrieval traffic, allow it, and verify its origin with care.
What they did on our site
Amazonbot's 2.9K reads have the shape of an indexer: broad passes across the site, repeated visits to the same pages, no clustering around the questions of the week. It read the homepage, the State of the Bots report and our product pages in roughly the proportions the site has them.
It sent no human back. A voice answer from Alexa carries no link, so there is nothing to click. The 47 AI referrals we recorded in the period came from ChatGPT, Claude and Perplexity. Amazon's reading is visible only in server logs, which is the general rule for AI visitors: they do not run JavaScript, so page analytics never records them.
Why commerce sites should care
Rufus answers "which one should I buy" inside the Amazon app. In 2025 Amazon tested Buy for Me, an agent that purchases from other brands' sites on a customer's behalf. For a brand or retailer, the practical questions are:
- Can Amazonbot read your product facts? Price, stock, specifications, compatibility and shipping as text in the HTML. A configurator that only exists in JavaScript is invisible to it.
- Are your comparison answers anywhere a crawler can find them? "Does X fit Y" and "X vs Y" questions are what shopping assistants are asked. An agent-only page in the bot-served sitemap answers them without touching the human site.
- Is the data consistent? The same price on your site and on third-party pages. Contradictions lower confidence in all of them.
- Can an agent act? Checkout for agents and machine payments are the next layer. Readable first, then usable.
TollBit's Bot Index tracks Amazonbot across 10,000+ sites, and an agent site serves it a clean, text-first version of every product page with the same facts the human page carries.
Frequently asked questions
Does blocking Amazonbot affect my Amazon listings or ads? No. Amazonbot is a web crawler. Your Seller Central listings, Amazon Ads and Buy Box standing run on Amazon's own systems and are unaffected by what your website serves to the crawler.
Is Amazonbot a training bot or an indexer? Amazon does not classify it either way. On our site it behaves like an indexer: broad, repeated, not question-driven. Decide how to treat it on what Amazon says it is for, which is improving Amazon services including Alexa's answers.
Should I treat Amzn-User like ChatGPT-User? Yes, as retrieval: allow it, serve it readable pages, measure what it fetches. Verify origin carefully, because the user agent is thinly documented and any token that sites tend to allow gets copied by scanners.