Amazonbot
Amazon's general crawler - and the only one of its three that Amazon says may be used to train AI models.
Reference reviewed by Sona
| Operator | Amazon |
|---|---|
| Powers | Alexa and Rufus answers, and Amazon AI model training |
| Purpose | Search & answers |
| User-agent token | Amazonbot |
| Respects robots.txt | Yes |
Amazonbot crawls web content to improve Amazon products and services, notably Alexa and the Rufus shopping assistant. Amazon's documentation also states it may be used to train AI models - which makes it the training-capable member of Amazon's crawler family and the one to focus on if training exposure is your concern.
Amazon splits its crawling across three tokens, and the split is cleaner than most vendors manage. Amazonbot is general-purpose and may train models. Amzn-SearchBot handles search experiences and explicitly does not crawl for generative AI training. Amzn-User handles live user-initiated fetches and also excludes training. If you want Alexa visibility without contributing training data, that is a supported configuration - allow Amzn-SearchBot, disallow Amazonbot.
Amazonbot respects the Robots Exclusion Protocol and honors user-agent directives. Amazon publishes IP address lists for each of its three crawlers separately rather than as one pool, so verification tooling needs the right list per token.
The commercial angle is worth weighing if you sell anything. Rufus answers product questions inside Amazon's own shopping surfaces, and Alexa reaches a large installed device base. For a retailer or brand, blocking Amazonbot has a more direct revenue shape than blocking a general-purpose training crawler does.
How Amazonbot behaves
- The only Amazon crawler documented as potentially feeding AI model training; its two siblings explicitly exclude it.
- Honors the Robots Exclusion Protocol and user-agent directives.
- Reverse DNS resolves under the Amazon-internal crawl.amazonbot.amazon domain, which impersonators cannot register.
- Has its own published IP list, distinct from the Amzn-SearchBot and Amzn-User lists.
Full user-agent string
Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; Amazonbot/0.1)
Allow Amazonbot
Your content can inform Alexa and Rufus answers across Amazon's device and shopping surfaces - commercially direct if you sell products.
User-agent: Amazonbot Allow: /
Block Amazonbot
This is Amazon's training-capable crawler, so blocking it is the way to opt out of Amazon AI training - and you can still allow Amzn-SearchBot for Alexa visibility.
User-agent: Amazonbot Disallow: /
How to verify Amazonbot
Reverse DNS + published IP ranges
Genuine Amazonbot traffic reverse-resolves under crawl.amazonbot.amazon - an unusual, Amazon-internal TLD that is itself a useful signal, since it cannot be registered by an impersonator. Amazon additionally publishes an IP address list for Amazonbot specifically, separate from the Amzn-SearchBot and Amzn-User lists, so point your verification at the right file rather than assuming one shared Amazon pool. Confirm reverse lookups with a matching forward lookup.
Documented reverse-DNS suffixes: .crawl.amazonbot.amazon
Commonly confused with Amazonbot
The distinction that matters: Amzn-SearchBot explicitly does not crawl for generative AI training, while Amazonbot may. Allowing the former and blocking the latter keeps Alexa visibility without training exposure.
Amzn-User handles live, user-initiated fetches and excludes training. Amazonbot is the scheduled bulk crawl.
bedrockbot is AWS-side: it crawls only URLs a Bedrock customer configured as a knowledge-base source. Amazonbot crawls on Amazon's own initiative.
Amazonbot FAQs
Which Amazon crawler is used for AI training?
Amazonbot. Amazon's documentation says it may be used to train AI models, while Amzn-SearchBot and Amzn-User are both documented as not crawling content for generative AI training.
Can I stay in Alexa results but out of Amazon's training data?
Yes. Allow Amzn-SearchBot, which handles search experiences and excludes training, and disallow Amazonbot, which is the training-capable crawler. Amazon documents the tokens separately, so the split works as written.
How do I verify Amazonbot traffic?
Reverse DNS on genuine requests resolves under crawl.amazonbot.amazon, an Amazon-internal domain nobody else can register. Amazon also publishes an Amazonbot-specific IP list - note that its three crawlers have three separate lists.
Does blocking Amazonbot affect my Amazon seller listings?
No. Amazonbot crawls the open web, not Amazon's own catalog. Your product listings inside Amazon come from your seller data, not from this crawler reading your website.
Can Amazonbot read your page right now?
Test any URL and see exactly what AI crawlers receive.