Are Duck.ai and Search Assist the same thing?
No. Search Assist (formerly DuckAssist) is the AI summary at the top of DuckDuckGo's results page; Duck.ai is a separate chat entry point. Below is what the official help pages let us verify (opened 2026-11-10; the help pages carry no date):
| Comparison | Search Assist | Duck.ai |
|---|---|---|
| Where | Search results page, above the results | A separate chat entry point; a Search Assist answer lets you continue with follow-up questions there |
| Source | Always links to one or two sources; looks for relevant material on the web, then writes a short AI summary | Unverified: the help pages do not say which models it uses or whether it searches the web |
| Official page | Help page Search Assist (see Sources below) | Unverified: we did not find a help page that describes it specifically |
So optimizing mostly targets Search Assist: its answers carry one or two source links, and the pages it picks can get clicks. Official help page: Search Assist.
Where do Search Assist's sources come from? Is Bing indexing a prerequisite?
DuckDuckGo's sources page says traditional web links and images come mostly from Bing, plus specialized sources such as Sportradar, crowd-sourced sites such as Wikipedia, and its own crawler DuckDuckBot and its own indexes. Which of these sources Search Assist picks pages from, the official pages do not say, and we have not verified it.
So the safe order is: get into Bing's index first, then let DuckDuckGo's own crawler in. For Bing setup see How to set up Bing Webmaster Tools: import, sitemap and IndexNow step by step; for how Bing relates to AI answers see Does ChatGPT search use Bing's index? Should you submit to Bing?. Sources page: DuckDuckGo Sources.
| Source | What the official sources page says | What you can do |
|---|---|---|
| Bing | Traditional web links and images come mostly from Bing | Submit a sitemap in Bing Webmaster Tools and confirm the pages are indexed |
| DuckDuckBot | Its own crawler, and its own indexes | Do not block it in robots.txt |
| Wikipedia and similar | Crowd-sourced sites | If your brand has an entry, keep it accurate |
| Sportradar and similar | Specialized sources, used for instant answers such as sports | Mostly irrelevant to an indie developer's product |
What is DuckAssistBot? How do I write robots.txt for it?
DuckAssistBot is the crawler DuckDuckGo uses to fetch pages in real time for AI answers. DuckDuckGo 的 DuckAssistBot 页 (checked 2026-11-10) says the following; DuckDuckBot's details are in the same table:
| Comparison | DuckAssistBot | DuckDuckBot |
|---|---|---|
| User-Agent | DuckAssistBot/1.2; (+http://duckduckgo.com/duckassistbot.html) | DuckDuckBot/1.1; (+http://duckduckgo.com/duckduckbot.html) |
| What it does | Fetches pages in real time to support AI answers with sources | The page says: improves search results; it is DuckDuckGo's web crawler |
| IP list | duckduckgo.com/duckassistbot.json | duckduckgo.com/duckduckbot.json |
| robots.txt | Disallow this User-Agent to opt out; takes effect after 72 hours; does not affect organic ranking | The page says it respects robots.txt |
| Used for training? | The page says: not used to train AI models in any way | The page does not say; unverified |
# Keep DuckDuckGo's own crawlers allowed (this is also what you get by saying nothing) User-agent: DuckDuckBot Allow: / User-agent: DuckAssistBot Allow: / # To opt out of AI answers only (DuckDuckGo: takes effect after 72 hours, does not affect organic ranking) # User-agent: DuckAssistBot # Disallow: /
Saying nothing means allowed, so you need no rule to be recommended. For other AI crawlers' names and purposes see AI crawler user agents: a side-by-side table; for general robots.txt syntax see How do you write robots.txt for AI crawlers?.
Do our four sites' robots.txt files name DuckAssistBot?
None names it, and none names DuckDuckBot, so the default User-agent: * rules apply. On 2026-11-10 we read the robots.txt of our four sites with curl, then fetched each home page with an ordinary User-Agent and with a DuckAssistBot User-Agent string:
| Site | robots.txt | Names DuckAssistBot | Names DuckDuckBot | Home page (ordinary / DuckAssistBot UA) |
|---|---|---|---|---|
| mochistamp.com | 200 | None | None | 200 / 200 |
| gammaslide.com | 200 | None | None | 200 / 200 |
| kagentx.com | 200 | None | None | 200 / 200 |
| pagent.kagentx.com | 200 | None | None | 200 / 200 |
How to read it: none of the four sites puts any special restriction on DuckDuckGo's two crawlers. This only checks the text of robots.txt and a User-Agent string; it is not the behavior of the real DuckAssistBot: our network is in mainland China, our IP is not in the ranges DuckDuckGo publishes, and we have not searched DuckDuckGo for these four sites, so whether Search Assist would pick them is unverified. The commands to check your own site:
# Does robots.txt mention the DuckDuckGo crawlers?
curl -s https://example.com/robots.txt | grep -i -B1 -A3 -E "duckassistbot|duckduckbot"
# Does the page answer the DuckAssistBot User-Agent?
curl -s -o /dev/null -w "%{http_code}\n" -A "Mozilla/5.0 (compatible; DuckAssistBot/1.2; +http://duckduckgo.com/duckassistbot.html)" https://example.com/page/
To confirm whether the real DuckAssistBot has visited, look at your server logs; see How do I check whether GPTBot and other AI crawlers have visited my site? for how.
What are the steps?
1. Check first whether Bing has indexed you
Search Bing for site:your-domain, or look at the indexed-page count in Bing Webmaster Tools. If nothing is indexed, submit first as described in How to set up Bing Webmaster Tools: import, sitemap and IndexNow step by step.
2. Make sure DuckDuckBot and DuckAssistBot are not blocked
Run the curl above; robots.txt should have no Disallow covering your content pages. Also glance at your CDN's block-AI-crawlers switch.
3. Put the body text directly in the HTML
Write the title as the buyer's question and give the answer and numbers in the first paragraph, so the summary has sentences ready to take.
4. Search DuckDuckGo with the buyer's questions a few times
Search 3 times and check whether a Search Assist answer appears above the results and whether you are among its source links. Answers differ each time, so count appearances. For how to record this see How do you measure AI visibility? Four metrics and a table you can follow.
5. Search the logs for DuckAssistBot
See which pages it has visited and what status codes it got.
Which sources the paid search engine Kagi (including Kagi Assistant) draws on is covered in How do I get Kagi to recommend my product? Kagibot and sources.
FAQ
What is DuckAssistBot's User-Agent?
DuckDuckGo's help page gives DuckAssistBot/1.2; (+http://duckduckgo.com/duckassistbot.html), with the IP list at duckduckgo.com/duckassistbot.json. It is the crawler that fetches pages in real time for AI answers with sources.
Will blocking DuckAssistBot drop me out of DuckDuckGo search?
No. The help page says opting out does not affect organic ranking or whether the site appears in results. You opt out by disallowing that User-Agent in robots.txt, and it takes effect after 72 hours.
Does DuckDuckGo train AI on my pages?
DuckDuckGo's DuckAssistBot page says this data is not used in any way to train AI models. That is its own statement and we cannot verify it independently.
Will Duck.ai recommend my product?
We have not verified this. DuckDuckGo's help pages do not say which models Duck.ai uses or whether it searches the web, so we cannot say where its answers come from. What we can confirm is that Search Assist answers carry one or two source links and that you can ask follow-ups in Duck.ai.
Should I get Bing indexing first?
Yes. DuckDuckGo's sources page says traditional web links come mostly from Bing, so a page Bing has not indexed will rarely appear in its web results. Whether Search Assist uses only those results, the official pages do not say, and we have not verified it.
Sources
- DuckDuckGo · Search Assist (Viewed 2026-11-10; the page carries no date)
- DuckDuckGo · DuckAssistBot (Viewed 2026-11-10; the page carries no date)
- DuckDuckGo · DuckDuckBot (Viewed 2026-11-10; the page carries no date)
- DuckDuckGo · Sources (Viewed 2026-11-10; the page carries no date)
- kagentx · kagentx.com/robots.txt (Read 2026-11-10)
- MochiStamp · mochistamp.com/robots.txt (Read 2026-11-10)
- Gamma AI Slide · gammaslide.com/robots.txt (Read 2026-11-10)
- Pagent · pagent.kagentx.com/robots.txt (Read 2026-11-10)
- kagentx evaluations: the data on the case study pages; each example gives its date.