Amazonbot vs. Amzn-SearchBot vs. Amzn-User: which should you block?

Amazon documents three crawler tokens on a single page, which is why a search for “amazonbot user agent” often surfaces more than one CrawlPact crawler page — they share a common Amazon origin, but each has a distinct purpose and, unlike some other operators’ crawler families, each has distinct robots.txt compliance behaviour too.

Which Amazon crawler should I block?

To keep content out of Amazon’s AI model training while staying eligible for Alexa and other Amazon search experiences, disallow Amazonbot and leave Amzn-SearchBot allowed — Amazon documents each token’s settings as independent. Amzn-User fetches pages for user requests and, Amazon says, may not follow all robots.txt directives.

What does each Amazon token do?

• Amazonbot — mixed use. Amazon’s documentation states it “is used to improve our products and services” and “may be used to train Amazon AI models.” This is the only one of the three associated with possible AI training.
• Amzn-SearchBot — search. Improves search experiences across Amazon products, including Alexa-eligible content. Amazon’s documentation states it is not used for generative AI model training.
• Amzn-User — user-triggered. Fetches a specific page in direct response to a person’s request, such as an Alexa query needing current information. Also not used for AI training.

What happens when you block each one?

• Disallow Amazonbot to opt out of Amazon’s general product-improvement crawling, including any possible AI-model-training use — this does not affect the other two tokens.
• Leave Amzn-SearchBot allowed if you want your pages eligible for Amazon’s search-style experiences. If you disallow it, note that Amazon’s documentation states this crawler otherwise falls back to whatever rule your robots.txt gives other search bots (a wildcard User-agent: * group, for example) when no Amzn-SearchBot-specific group exists — a blanket allow for “all search crawlers” already covers it.
• A Disallow rule for Amzn-User is not a reliable way to stop it. Amazon’s own documentation states this token “may not follow all robots.txt directives” because its requests are triggered by a real person’s action, not a bulk crawl — the same design choice Perplexity makes for its own user-triggered fetcher (see PerplexityBot vs. Perplexity-User). If you need to prevent this specific fetch, robots.txt alone will not guarantee it; see /limitations for what a robots.txt rule can and cannot guarantee.

Does an Amazonbot rule cover Amazon’s other crawlers?

Site owners sometimes write a single Amazonbot rule expecting it to cover Amazon’s crawling broadly, then are surprised when Amzn-SearchBot or Amzn-User activity continues (or, in Amzn-User’s case, continues regardless of what robots.txt says at all). Each token is matched independently, and one of the three has documented, limited robots.txt compliance in the first place. CrawlPact’s AI crawler checker evaluates all three against your robots.txt so any unintended gap between them is visible before it matters.