Free tool

Non-Commodity Score

AI search does not cite pages that say what everyone else already says. This scores any page 0 to 100 on how far it is from commodity content: first-hand experience, hard numbers, and entity grounding, each graded with quoted evidence rather than a black-box verdict. Two scores free, then unlock with your email.

Two free scores, then unlock with your email. Scoring takes 10 to 20 seconds; the experiential grade is an AI judgement with quoted evidence.

What it measures

The score is a composite of four vectors. Experiential evidence is graded by an AI judge against a rubric (first-hand actions, specific conditions, outcomes with numbers) and must quote the passages that justify its grade, so you can check its working. Empirical telemetry counts unit-bearing values, precise numerics and data-table rows as a saturating density. Entity connectivity parses every JSON-LD block and counts entities that resolve to external authorities such as Wikidata, plus outbound links to authoritative domains. Information gain, the fourth vector, needs the full ranking corpus for your target query, so this free build reports it as not measured and renormalises the composite rather than guessing.

Why "uncalibrated: directional only"

The weights behind the composite are a starting prior, not a fitted model. Until they are calibrated against pages with observed outcomes (cited in AI Overviews, holding rankings), an honest tool reports a diagnostic profile, not a pass or fail verdict. The story of why we refuse to ship invented precision, including the teardown of our own first design, is in the build log.

One honest limit: the tool reads the raw page source, so content and schema injected only by client-side JavaScript are not visible to it.