Paste any paragraph. Get a 0–100 liftability score showing exactly what ChatGPT, Perplexity, and Google AI see when they decide whether to cite you.
100% in your browser · your text never leaves your device
An answer capsule is the first 40–60 words of a section — the chunk an AI engine lifts when it cites you. This tool scores how citable yours is across 5 weighted dimensions, with inline highlights so you can see exactly what the engine reads.
New here? Read how to show up in AI search · or the GEO field guide
LumenGEO — the SaaS I built — runs live AI queries across ChatGPT, Perplexity, and Google to show you your actual citation share and the competitors beating you right now.
AI search engines don't read your article — they lift passages. The first 40–60 words of each section are the unit of retrieval. Four properties make a passage the one the engine picks over everyone else's.
The first sentence must answer the question. Not "in this section" — the actual answer. Engines grab the opening and stop. Preamble means you lose the citation.
40–60 words is the sweet spot. Under 20 is too thin to cite confidently. Over 85 and the engine has to crop — usually badly. Structure your section breaks around this window.
Never start a citable passage with "This works because…" or "However, it also…" — the engine pulled that passage in isolation. There's no "this." Use the noun, not the pronoun.
"Stripe charges 2.9% + 30¢" is 6× more citable than "providers offer flexible pricing." Named entities, numbers, and percentages are why engines cite some passages and skip others.
These four rules come from the same 87-experiment dataset behind the GEO field notes and the operator's playbook. They're not hypotheses — they're patterns measured on a live site generating 135,700+ Copilot citations in 5 months. The scoring weights in this tool reflect relative impact across those experiments. Lead directness accounts for 28 points because a buried lede is the single highest-frequency failure mode: it appeared in 61 of the 87 failing sections analyzed.
One thing this tool can't measure: whether your page is indexed and being retrieved at all. A perfect capsule on an unindexed page earns zero citations. That's what the GEO Readiness Scanner covers — run both if you're doing a full audit.
If you want me to audit your full content set against live AI responses — finding the exact queries where you're losing citations to competitors — that's LumenGEO or a direct engagement.