FanoutFox: AI Answer & GEO Inspector

See how ChatGPT and Perplexity pick and cite sources: fan-out, citations, brand mentions, trust notes. Reads only your own session.

See exactly how ChatGPT and Perplexity pick the sources behind an answer, and whether your brand made it in. FanoutFox reads your own AI search session and shows you the searches the engine actually ran, every page it fetched, which of those earned a citation, and which brands got mentioned in the text people read. Everything stays in your browser. It's free and there's nothing to sign up for. WHAT PROBLEM IT SOLVES ChatGPT doesn't run your keyword. It rewrites your question into its own queries, fetches pages you never see, and works out on its own which brands get mentioned. Classic rank tracking measures a results page your buyers increasingly never open. FanoutFox opens the box, whether the job that day is GEO, ChatGPT and Perplexity SEO, or AI answer rank tracking. HOW IT WORKS Install, pin the fox, and on first run create a workspace with your brand name, your domains and a few competitors. Then browse ChatGPT or Perplexity as you normally would, and click the fox when an answer that searched the web finishes. Captures save locally as you go, and sample data built from genuine captures lets you explore every view first. WHAT YOU GET • The fan-out. Every search the engine actually ran, which is almost never your keyword. Write for the query it really runs. • Sources versus citations. Dozens of pages get fetched, a handful get cited, and you see which side of the line you're on. • Mentions. The exact sentence that mentioned you or a rival, because a mention sits in the words buyers actually read while a citation is a link most never click. • Conversion. The share of each domain's reads that earned a citation. A page read often and never quoted is a content problem rather than a crawling one. • The runner-up ledger. Per claim, who won the citation and who came second, including you. • The trail. The ordered path from question to answer, with the model's own words about why it chose a source. • Shopping tiers and Reddit intel. Product cards versus text mentions, with a flag when a rival got the card, and the threads that provably fed the answer. PERPLEXITY, AND COMPARING ENGINES Perplexity capture adds two signals of its own, the per-source trust notes it returns and how its router classified your question. The dashboard scopes every section to one engine at a time, labels the features only one engine ships, and switches off the ones an engine can't feed rather than showing empty states. The Compare view then lines the same prompt up across both engines, and that comparison is the closest thing AI answers have to rank tracking. TRACKING CHANGE OVER TIME Pin a capture as a baseline, re-run the prompt later, and diff who entered, who dropped out, and who moved from mentioned to cited. The representation audit scores the stored summary the engine holds for each of your pages. Personalised answers get flagged, since your placement there is a mirror rather than the market. Exports cover CSV, Markdown, JSON and a one page brief that turns losses and gap domains into a prioritised fix list. PRIVACY Most SEO tools want your data on their servers. This one doesn't have a server. It reads only your own sessions, saves captures in your browser's own storage, and uploads nothing. On ChatGPT the session token your browser already holds is used once to fetch your own conversation, then discarded. On Perplexity no token is read at all. Two optional lookups, a domain rating and Reddit thread stats, are the only outbound calls and both switch off. Uninstall it and every trace goes with it. WHO IT'S FOR SEO and GEO practitioners, content and PR teams, agencies reporting on client visibility, and founders who want to know whether AI answers recommend them or a competitor. WHY IT'S FREE It reads undocumented fields the platforms change often. When a field moves, a card goes quiet and says so, and a fix ships. A tool on ground that shifts that fast has no business charging a subscription. RECENT UPDATES • v1.1.8. Features an engine can't feed switch off with a label instead of posing as empty results, and the retired ChatGPT retrieval pipe column says why. • v1.1.5 to v1.1.7. Perplexity's router card and trust notes on every surface, an Engine signals section, and engine-only badges down to the popup row. • v1.1.4. The conversion metric and a citation type tripwire. The full changelog, the research it's built on and a manual download live at fanoutfox.com.