In short. Amzn-User retrieves live web information on behalf of users, with a documented exception to robots.txt compliance for user-initiated actions.

Definition: Amzn-User is Amazon's user agent for retrieving live web information on behalf of users, including answers to Alexa questions that need current information. As of 17 October 2026, Amazon documents an exception to robots.txt compliance for these user-initiated actions.

Key facts

Amazon's About Amazonbot documentation describes Amzn-User fetching web information when a customer asks a question.
Amazon publishes a separate live retrieval IP address list containing an ipv4Prefix field for each listed address.
The IETF's Robots Exclusion Protocol treats robots.txt rules as crawler instructions and recommends authentication for access control.
The IETF's HTTP Semantics specification defines 403 Forbidden as a server understanding a request and refusing to fulfill it.

When does Amzn-User retrieve a page?

Amazon's Amzn-User description gives the example of an Alexa question requiring up-to-date information. A customer action can prompt a live fetch, meaning a request for information from the web while handling that action.

The same documentation assigns separate purposes to Amazon's agents. Amzn-SearchBot supports search experiences in Amazon products, while Amazonbot improves products and services and may collect content for Amazon model training. Amazon excludes model-training crawling from Amzn-User's purpose. Each agent has independent settings, so an Amazonbot rule alone leaves Amzn-User's setting separate.

How can a website identify these requests?

A user-agent string is the identifier a client sends with its request. Amazon's documentation supplies this example for Amzn-User:

Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; Amzn-User/0.1) Chrome/W.X.Y.Z Safari/537.36

The W.X.Y.Z component appears as a placeholder in the published example. For an address check, a site operator can compare a request's source IP address with Amazon's Amazonbot Live Crawl IP addresses page. Its list uses the fields prefixes and ipv4Prefix. This provides an address reference specifically linked from the Amzn-User documentation.

How should access restrictions be handled?

A robots.txt file publishes instructions for automated clients. Amazon's robots documentation describes support for User-agent, Allow and Disallow during automated crawling. A group expressing a site-wide restriction for Amzn-User takes this form:

User-agent: Amzn-User
Disallow: /

Amazon explicitly allows an exception for actions initiated by users: Amzn-User may leave some robots.txt directives unfollowed. The documentation leaves the affected directives unspecified.

For content requiring restricted access, the IETF's robots protocol security guidance recommends an application security measure such as HTTP authentication. Publishing a path in robots.txt also exposes that path publicly.

A server's response expresses the access decision. Under HTTP Semantics, 401 Unauthorized indicates missing valid authentication credentials and requires a WWW-Authenticate challenge. 403 Forbidden expresses refusal to fulfill the request. These response meanings provide a basis for checking whether the server actually serves a protected page.

FAQ

Can I rely on robots.txt to prevent every Amzn-User fetch?

Amazon's Amzn-User policy permits exceptions for user-initiated actions. For restricted pages, the IETF's security guidance recommends access controls such as HTTP authentication.

Does Amazon describe Amzn-User as a training crawler?

Amazon's agent descriptions exclude model-training crawling from Amzn-User's purpose. The same page identifies Amazonbot as an agent whose collection may support Amazon model training.

Sources

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