DuckAssistBot: real-time reads for DuckDuckGo answers
DuckAssistBot is the crawler DuckDuckGo uses to read pages in real time for its AI-assisted answers. Those answers cite their sources prominently, so a fetch can turn into a link back to you.
DuckDuckGo makes two promises about it that few operators state so plainly. The data is not used to train AI models, and opting out does not affect your organic rankings.
DuckAssistBot/1.2; (+http://duckduckgo.com/duckassistbot.html)
- Operator
- DuckDuckGo
- Type
- Live fetch for a user
- robots.txt token
DuckAssistBot- Obeys robots.txt
- Yes, by the operator's documentation. DuckDuckGo says an opt-out takes effect after 72 hours
- Official docs
- duckduckgo.com
Why DuckAssistBot is an easy call for most sites
Many AI crawlers force a trade between visibility and training. This one does not, because DuckDuckGo says it uses none of the data for training. So the only question is whether you want to be a possible source in its AI-assisted answers.
Also, the opt-out is narrow. DuckDuckGo says blocking the agent does not affect whether your site appears in its regular search results. In other words, you are choosing about the AI answer box only.
| Question | Answer |
|---|---|
| Used for AI training? | No |
| Affects organic rankings? | No |
| Obeys robots.txt? | Yes, after 72 hours |
| IP list? | Yes, on the help page and in JSON |
How to opt out of DuckAssistBot
DuckDuckGo asks publishers to use robots.txt. Add a group for the token, and the crawler stops after 72 hours, which is longer than the roughly one day that OpenAI or Perplexity mention.
User-agent: DuckAssistBot
Disallow: /If something goes wrong with the opt-out, DuckDuckGo lists a contact address for crawling problems on its help page.
There is no need for a server rule, since the real crawler follows robots.txt. Likewise, Pagereckon Pro's one-click blocking is meant for bots that crawl hard and give nothing back, such as Bytespider, and does not cover this one.
What opting out costs
Your pages can no longer serve as cited sources in DuckDuckGo's AI-assisted answers. Each such citation is a chance for a click, so the loss is real, if small for most sites.
On the other hand, nothing else moves. Your organic placement on DuckDuckGo stays as it was, per DuckDuckGo, and none of your content was going into model training to begin with.
Verify a DuckAssistBot request
DuckDuckGo prints the crawler's source IP addresses on its help page and offers the same list as JSON at duckduckgo.com/duckassistbot.json. Match the IP from your log against that file.
If the address is not on the list, the request did not come from DuckDuckGo, whatever its user agent says.
DuckAssistBot in Pagereckon
Pagereckon files the agent under live reads, in a Duck.ai row, because it fetches in real time while an answer is being built. That puts it next to ChatGPT-User and Perplexity-User rather than next to the training crawlers.
As a result, a rise in these reads means DuckDuckGo is using your pages in more answers. See the full split in the AI crawler checker, and compare it with Perplexity-User, which works in a similar way.
DuckAssistBot questions
Does DuckAssistBot train AI models?
No. DuckDuckGo says the data is not used in any way to train AI models.
Will blocking it hurt my DuckDuckGo rankings?
No. DuckDuckGo says opting out does not impact organic search rankings or whether your site appears in results.
How long does the opt-out take?
DuckDuckGo says the change takes effect after 72 hours.
Can I see its reads for free?
Yes. Download the free plugin and they appear under live reads.