Citation tracking — the pages AI trusts
AI answers are assembled from cited pages, and those pages are a market you can enter. CiteGraph captures every citation across every run, folds them into a ranked source list — graded by how substantive the content is and how gettable a mention would be — and flags the citation targets: pages that answer the questions you lose.
How it works
Captured per answer, folded per domain
Every cited URL from every run is stored, deduplicated to root sources, and ranked by the share of answers citing it. Redirect-wrapped grounding URLs are resolved to the real source.
Graded, not just counted
Each source is typed (roundup, review platform, community, reference, press) and graded: substantive coverage, thin mention, or noise — because a citation from a real comparison is worth chasing and a keyword-stuffed page is not.
Turned into targets
Sources cited on questions where you were never named become citation targets, ranked by how many lost questions each answers. The pitch angle for each is drafted in Plays.
In depth
Gettability is stated up front
A niche roundup that updates monthly is easy; Wikipedia is a long shot. Every source carries an easy / worth trying / long shot grade so effort lands where it can work.
Evidence folded in
Each source keeps the actual answers that cited it — the question, the engine, the excerpt — so you can read why this page keeps winning before you decide to pitch it.
Co-citation mapped
Sources that appear together reveal the answer's skeleton: get into one member of a co-cited cluster and you're suddenly adjacent to the rest.
What we won't sell: placement. Some tools operate backlink marketplaces next to their citation data. We think that's a conflict of interest wearing a dashboard — the targets are yours to earn, and we tell you how.
Common questions
Which pages do AI engines actually cite?+
In our scans: roundups and listicles, review platforms, community threads, reference pages, and product docs — in proportions that vary sharply by engine and category. Your scan shows your category's real mix, with counts.
What is a citation target?+
A third-party page cited on buyer questions where you were never named. Getting fairly mentioned there puts you inside answers you currently lose. Your scan ranks them by how many lost questions each one answers.
Do you resolve redirect URLs?+
Yes — grounding APIs often wrap citations in redirect hosts; we resolve pairs so the ranked source is the page a human would credit, not a proxy domain.
See it on your own domain — first dossier in about a minute.