Every ArticleGrade score is a 0–100 number built from the same rubric — four buyer-legible signals, six underlying dimensions, twenty-six gates — read by two independent models. This is how it's computed, and where its limits are.
Thin, unsupported or visibly machine-written. The kind of page that drags a domain down — block it.
Publishable in places, but with real gaps in sourcing, depth or structure. Remediate before it ships.
Specific, sourced and human enough to defend. 70 is the default gate; you can set your own.
The headline number is the stricter of the two model reads on overall quality. You choose the threshold your pipeline gates on — the score is a measurement, not a mandate.
The four signals are how we report the score; underneath, six dimensions and 26 pass/fail gates do the work. Each signal rolls up the dimensions below it.
Whether the page demonstrates experience, expertise, authority and trust — author signals, source authority, and the markers Google's quality raters look for.
The tells of unedited generation — hedging, filler, repetition, template leakage and a flat machine register that readers and classifiers both notice.
Whether claims are specific, sourced and verifiable, and whether the piece has enough substance to actually answer the query.
Title, meta, heading hierarchy, internal links, formatting and CMS artifacts — the technical layer that frames everything else.
We split the judgement across two models so neither marks its own homework — and so we can screen cheaply before spending on depth.
Judges whether the content is actually good — E-E-A-T, factual depth, sourcing and AI-artifacts. Its read drives the headline score and the remediation loop.
Judges only the technical layer — title, meta, headings, links, formatting. It explicitly does not fact-check, so it can't "correct" the page from stale training data.
A quality score is a strong signal, not an oracle. We'd rather you know exactly what it is — and isn't.
Google's E-E-A-T is a rater framework, not a published ranking number. Our E-E-A-T score is a defensible proxy built from the same signals — directional, not a readout of Google's internal score.
One weak page won't sink a domain — Google's quality signal is about the share of unhelpful content. We lower that share by catching weak pages before they publish, not by promising a single page is decisive.
The score reflects the rendered URL as Google indexes it — including the chrome, schema and template your CMS adds. That's the version that actually ranks.
The rubric was calibrated on hundreds of real audits and is refined as Google's guidance shifts. A score is a snapshot against today's bar, not a permanent grade.
Grade any URL free and read the score, the dimensions and the issues for yourself.