Documentation
Everything about how GEO Tracker measures AI visibility
Read how scores are calculated, what the engines actually return, and the trade-offs we made along the way. We try to be honest about what the product does and what it doesn't.
Concepts
- How the GEO Score is calculatedThe exact formula we use to turn raw mentions into a single 0–100 visibility number, with worked examples.
- What AI engines we trackChatGPT, Perplexity, and Google AI Mode — how each one differs, what they return, and why we weight them the way we do.
- Google AI Mode — what it is, how we track itThe dedicated AI tab on google.com is a different surface from classic AI Overviews. Here's what we measure and how.
- AI citation vs. source URL — what counts as whichTwo different signals AI engines emit, and why we treat them separately.
- Citation Source Intelligence — what it is and the three layersWhy we don't treat AI citations as undifferentiated URLs — and the three-layer architecture (extract, enrich, act) that turns a list of cited pages into a queue of paste-ready actions.
- Citation actionability — Live, Limited, Frozen, ManualWhy we classify every cited URL into four states before generating a draft, and what each state means for the action you can take.
Features
- Manual, pulse, and deep scansThe three scan rhythms that drive every dashboard number — what each one does, when it runs, and what counts toward your tier limits.
- Content audit — what we crawl, score, and generateThe multi-page audit that grades how AI-discoverable your site is, plus the JSON-LD and FAQ schema generators that fix the gaps.
- Insights feed and weekly Action PlanHow the 'What changed' timeline detects movement and turns it into specific actions you can take this week.
- Citation excerpt search — Reports tabFind every AI answer that mentioned a term across the last 90 days, with two-color highlighting and day/scan-run grouping.
- Citation Source Intelligence — Citations tabWhere AI engines pull citations from for your queries, whether you can act on each source, and paste-ready drafts (Reddit, HN, email pitch, GitHub PR, podcast) for the ones that are open.
- Suggested queries — how the suggest engine worksHow we propose monitored questions for your domain — multi-page grounding, category presets, and the provenance badge that tells you which is which.