In short. Amazonbot is Amazon's web crawler with published robots.txt controls and a noarchive instruction for excluding pages from model training.

Definition: Amazonbot is Amazon's web crawler for improving its products and services, including collecting information that may train Amazon AI models. As of 17 October 2026, Amazon documents controls through robots.txt and page-level robots meta tags.

Key facts

Amazon's About AmazonBot documentation separates Amazonbot from Amzn-SearchBot and Amzn-User, with independent settings for each agent.
Amazon publishes an Amazonbot IP address list containing a prefixes array with ipv4Prefix entries.
The IETF's Robots Exclusion Protocol specification places crawling rules in /robots.txt at the service's top-level path.

How does Amazonbot differ from Amazon's other agents?

A user agent identifies the software making a web request. Amazon's crawler documentation gives Amazonbot/0.1 as the Amazonbot token in its example user-agent string. Its stated purpose includes improving the accuracy of information supplied to customers and potentially training Amazon AI models.

Amzn-SearchBot serves search experiences in Amazon products, including Alexa. Amazon says it excludes model-training collection from this agent's purpose. When robots.txt omits Amzn-SearchBot but allows other search bots, it follows the directives given to those other search bots.

Amzn-User supports user actions, such as fetching current web information to answer an Alexa query. Amazon says user-initiated actions may bypass some robots.txt directives. The documentation also excludes model-training collection from this agent's purpose. These separate roles make the agent name relevant when choosing a control.

How can a website block Amazonbot crawling?

A robots.txt file publishes instructions for crawlers. Under Amazon's published robots.txt policy, automated crawling honors User-agent, Allow and Disallow directives. The following example applies a site-wide disallow rule to Amazonbot:

User-agent: Amazonbot
Disallow: /

This example adapts the named-agent and root-path disallow syntax shown in the IETF specification. Amazon reads rules at the host level: example.com/robots.txt and site.example.com/robots.txt control their respective hosts. A domain with several hosts therefore needs rules on each relevant host.

Amazon says setting changes may take approximately 24 hours to appear in its systems. Its policy also permits a cached robots.txt copy from the last 30 days. If the file cannot be fetched, the agents behave as though it is absent.

The Robots Exclusion Protocol describes these rules as crawler instructions and calls for separate security measures, such as HTTP authentication, to protect access to content.

Which page-level opt-out controls does Amazon publish?

A robots meta tag is an HTML element carrying page-level instructions. Amazon's documentation on page controls assigns noarchive the meaning of excluding the page from model training. An HTML implementation matching that documented instruction is:

<meta name="robots" content="noarchive">

The same documentation assigns noindex and none the meaning of excluding a page from indexing. It also says the agents respect rel="nofollow" on links. Amazon describes these page controls as applying when its agents access web pages, so the training opt-out belongs in the page's HTML where the agent can read it.

FAQ

Does blocking Amazonbot also block Amzn-SearchBot?

Amazon's crawler documentation says each agent's setting is independent. To control search crawling, address Amzn-SearchBot separately.

Can I slow Amazonbot with crawl-delay?

Amazon says its listed agents do not support crawl-delay. Its robots.txt policy instead documents support for user-agent selection and allow/disallow rules.

Which tag excludes a page from Amazon model training?

Amazon documents noarchive for that purpose. Its meanings for noindex and none concern page indexing.

Sources

Check how a page reads to answer engines. Open the GEO score analyzer