In short. DuckAssistBot fetches pages for DuckDuckGo's AI-assisted answers and supports a robots.txt opt-out that takes effect after 72 hours.

Definition: DuckAssistBot is DuckDuckGo Search's web crawler, a program that fetches web pages, for AI-assisted answers. As of 17 October 2026, DuckDuckGo documents real-time crawling and a publisher opt-out through robots.txt.

Key facts

DuckDuckGo's DuckAssistBot help page states that data collected by this crawler is not used to train AI models.
The Search Assist documentation describes brief answers based on web content, with direct links to one or two sources.
DuckDuckGo separately documents DuckDuckBot as a crawler whose job is to improve its search results.

Which answer feature uses the crawled pages?

Search Assist is an optional feature within DuckDuckGo's search results. DuckDuckGo's Search Assist help page explains that the service scans the web for relevant content and uses AI-powered natural language technology to produce a brief answer from the information found.

The same documentation says that responses link directly to their sources. Readers can select “More” to expand an answer and see additional detail. DuckAssistBot supplies real-time page retrieval for these AI-assisted answers, as described in DuckDuckGo's crawler documentation.

How can a website identify DuckAssistBot requests?

A user-agent string identifies the crawler in its requests. DuckDuckGo's DuckAssistBot documentation publishes this exact string:

DuckAssistBot/1.2; (+http://duckduckgo.com/duckassistbot.html)

The same page lists the IP addresses from which the crawler operates and links to a machine-readable JSON list. For a review of incoming requests, the published string and IP list provide reference values for comparison with the website's request records.

How does a publisher opt out?

A robots.txt file supplies crawling rules for a website. The IETF's Robots Exclusion Protocol, RFC 9309, places the file at the service's top-level /robots.txt path. Its rules associate a crawler name with allowed or disallowed paths.

The following group expresses a site-wide disallow rule for DuckAssistBot using that protocol's syntax:

User-agent: DuckAssistBot
Disallow: /

DuckDuckGo's publisher instructions say that disallowing this user agent takes effect after 72 hours, when DuckAssistBot stops crawling the site. The opt-out concerns being a potential source for AI-assisted answers. According to the same instructions, opting out leaves organic search rankings and whether the website appears in search results unaffected.

FAQ

Does allowing DuckAssistBot ensure that a page appears in an answer?

DuckDuckGo's crawler help page describes participating publishers as potential sources for AI-assisted answers. Allowing access leaves the page available for that use.

Where can publishers report an opt-out problem?

The same help page directs publishers with opt-out problems to crawling@duckduckgo.com. It specifies a 72-hour delay for an updated disallow rule to take effect.

Sources

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