Free tool · Free, email required
AI citation source radar: which sites the answers cite
How it works
Free, asks for an email. About twenty seconds.
Enter your site
We read it and write a short description and the keywords your buyers would use, which you can check before anything runs.
Three buying questions go to Perplexity
Each is a live answer that searches the web, and every page it cites is kept and grouped by the site it came from.
Read who supplies the answers
A leaderboard of sites by share of citations, the kind of page each one contributes, and whether your domain appears at all.
Why concentration is the number that matters
People assume AI answers draw on some vast, evenly-weighted read of the web. They do not. An engine retrieves a small number of pages per question, and across a category the same hosts keep getting retrieved. A handful of sites end up supplying most of what buyers hear about your market.
That is bad news and good news in the same fact. Bad, because you cannot fix it by writing more on your own blog. Good, because the target list is short and knowable. If five hosts produce half your category's answers, you have five things to work on, not the whole internet.
Read the page types, not just the names
The leaderboard labels each host by the kind of page it mostly contributes, and that label decides what you can do about it.
Best-of lists and comparison pages are the highest-value target: they exist to enumerate options, and adding one is usually an email to the author rather than a content project. Review directories like G2 and Capterra cannot be written into — you need a profile and reviews, which is a slower and more expensive fix. Competitor pages are their content winning, and the response is your version of the same argument on your own domain. Discussion threads are the ones people most want to game and should not; the durable play is being the page those threads link to.
What this run does not tell you
Three questions on one engine, once. That is enough to see the shape of your category and not enough to make a quarter's plan from. A host with a big share here might be a genuine category authority, or it might have got lucky on the specific phrasing we used. The way to tell those apart is repetition across more questions, more engines, and more weeks, which is what the paid product does and why this page is free without being a trial.
Every day, not once
AI citation tracking
The free scan reads three questions once. CueScout keeps every page the engines cite for all your questions, checks whether you are on each, and sorts them by what you can win.
Keep going
All free toolsQuestions
Which pages does an AI answer cite?
The pages an AI engine actually read to build its answer. Ask ChatGPT or Perplexity a buying question and it retrieves a handful of pages, reads them, and summarises. Across a category the same hosts turn up again and again, and that concentration is what this tool measures.
How does this differ from a visibility checker?
A visibility checker answers "am I named". This answers "who is doing the naming". The second question is more useful when the answer to the first is no, because it tells you where the answer is coming from instead of just confirming you are not in it.
Why do you ask for my email, and what do you keep?
Each run spends real money: three questions put to an AI engine that searches the web, from a page anyone can use without logging in, so the email is friction against scripts. We keep the scan with your email, the brand and domain, the questions asked, the pages cited and the network it came from. That record is also what lets you carry the results into an account if you sign up later, without us re-running the scan. The results appear on this page; they are not emailed.
Three questions is not many.
No, it is not, and the concentration figure from three questions is directional rather than precise. It is enough to see whether your category is dominated by a handful of hosts, which is the thing most people have never actually looked at. The full prompt list, tracked daily on the engines your plan covers, is the paid product, and it exists because three questions once is genuinely not enough to make decisions on.
My competitors are all over the leaderboard and I am nowhere. Now what?
Look at what kind of pages they are. If the top hosts are best-of lists and directories, the job is getting a row added, which is outreach, not writing. If they are competitor blog posts, the job is publishing the better version on your own domain. If they are Reddit and Stack Overflow threads, the job is being the answer those threads link to. Different problems, different fixes, and the leaderboard is what tells you which one you have.
Do the results change between runs?
Yes. Engines re-search each time and the underlying results move. We cache a result for 24 hours per brand, so running it twice in a day gives you the same answer, mostly so a link doing the rounds does not cost us one scan per visitor. Across weeks the cited sites do shift, and watching that shift is the thing worth paying for.
Is the discussion-thread number an invitation to go post on Reddit?
No, and it is worth saying plainly. Threads appear because engines cite them. CueScout has no mechanism to post anywhere and does not recommend replying in threads to manufacture citations. That approach gets accounts banned, which we know because it happened to ours, four times, which is why the product works the way it does now.
Track it every day instead of once
CueScout runs these checks on a schedule and turns the gaps into To-dos.