See which pages AI search engines actually cite

AI answer engines do not answer from a single ranking position — they retrieve several sources per question and synthesise them, and they reformulate the user's question into multiple internal queries along the way. Retrieval overlap measures which pages survive that process: the pages that keep being retrieved across many phrasings of the same topic. Those recurring sources are the ones most likely to end up in an AI answer, which is why overlap is a better proxy for AI visibility than rank alone.

There is no subscription. A signed-in run across four AI engines costs 135 credits, and you are only charged when a run produces a result — a failed run returns the hold automatically. The 500 credits you get on signup cover 3 runs. Credits do not expire.

Anonymous runs are Google-SERP-only. A signed-in run also queries Claude, Perplexity, Gemini, and ChatGPT and records the pages each engine actually cited for every query. That unlocks the AI citation breakdown, per-engine citation tracing you can verify yourself, a cross-engine agreement map, and the "cited but not ranking" view.

These are pages an AI engine cited when answering your cluster that never appear in Google's top 20 for any query in it. They are invisible to conventional rank tracking, yet they are actively shaping AI answers. The tool lists them, cross-checks the deterministic signals behind them (platform mix, UGC share, cited-query coverage), and can analyse the common pattern explaining why they earn citations without Google visibility.

On Google-only runs it blends query coverage across the cluster (weighted most heavily), prominence on user-generated platforms, and best achieved position. When AI engines are active, the score shifts to weight the share of queries where the page was genuinely cited, so pages that AI actually uses rise above pages that merely rank.

It measures how much each pair of engines agrees on which pages recur, as an overlap percentage, and lists the specific pages behind it. A page that several independent engines converge on is the strongest possible citation target for a topic — earning presence there influences multiple answer engines at once rather than just one.

Yes, and that is deliberate. Signed-in reports include a per-engine trace listing the exact pages each engine cited, grouped by query, as clickable links. You can open the engine, ask the same question, and confirm the citations directly. Every engine also reports its own coverage, so an engine that errored is shown as failed rather than silently dropping out of the numbers.

A rank tracker answers "where does my page sit for this keyword". AI Search Visibility answers "which pages does search keep pulling for this whole topic, and am I one of them". It works at cluster level rather than keyword level, counts recurrence rather than position, includes user-generated sources that rank trackers deprioritise, and — signed in — measures real AI citations rather than inferring them from rankings.