Skip to content
traze

New · free tool · no login

Content Scorer

Paste a URL and a keyword. Within three seconds you get a citability score for AI engines, with the concrete change per component.

Try:

No login. We fetch the page, assess it in memory and throw the content away afterwards.Full result, no email needed.

What the check looks at · 5 components

What the Content Scorer assesses

An AI answer cites pages that answer a question quickly, concretely and verifiably. The scorer assesses exactly those properties — the same criteria Traze checks every page it writes against before it goes live. Counts (words, headings, lists, schema) are measured; the six judgements come from our decision model, with their certainty attached.

  1. 01

    Direct answer

    Is the answer to the main question in the first paragraph, in two or three sentences that can be quoted on their own? Or does a reader — and a language model — have to get through an introduction first?

  2. 02

    Entities and context

    Does the page name the products, places, brands and concepts that belong to the topic, and always by the same name? Engines tie your page to a topic through those entities.

  3. 03

    Structure

    Headings that are questions, short paragraphs, lists and tables where they fit, and an FAQ. Structure decides which pieces of text can be lifted out as an answer.

  4. 04

    Evidence and provenance

    Numbers with a source, an author with expertise, a last-updated date and schema markup. Engines prefer pages whose reliability they can estimate.

  5. 05

    Match with the keyword

    Does the page cover the intent behind the keyword — informational, comparison or purchase — and the sub-questions people usually ask alongside it? A page that answers half the question is rarely cited.

Why it matters · the short version

Why citability is the new ranking

AI engines don't pick pages. They pick passages.

A classic search engine ranks pages and lets the reader choose. An AI engine reads a handful of pages, lifts out the passages that answer a question and builds one answer from them, with a citation for whoever gets quoted. Whether you appear in that answer depends less on your authority as a domain and more on how citable the text itself is.

That is good news for anyone who takes it seriously. Citability is measurable and improvable without link building or budget: a direct answer at the top, clear entities, a structure that makes pieces of text stand alone, and evidence an engine can verify. The same changes make the page better for Google and for the reader.

The Content Scorer gives you a score per component and the concrete fix. Inside Traze this check runs on every page it writes, before publication, and again as soon as an engine stops citing the page. The free version shows how the check works on one page of your choice.

Methodology · what is what

What we decide, measure and estimate

Every number in the result belongs to one of these three. The report itself says which, per item, so you know what to lean on.

Assessed by our decision model

A judgement with a probability behind it: certain, likely or uncertain. Not a fact, but traceable — the same input gives the same judgement, and the certainty is always shown.

The six judgements about the page (direct answer, entity, citability, readability, provenance, summarisability) and which change per area helps most.

Measured

What we literally fetched or counted: HTTP status, HTML, robots.txt, sitemaps, feeds and search results. If it is not there, we say so.

Everything we count on the page: words, paragraphs, headings, lists, tables, links, images without alt text, schema types, author and date signals.

Estimated

Models and rules of thumb: search volumes, click rates, expected return. Good enough to prioritise with, not to budget on.

The final score: a fixed weighting over the six judgements, so one number stays comparable between pages.

Simulated. When no decision model, provider or cluster is available — or the daily ceiling for the free tools is reached — the tool keeps running on deterministic example data. That is labelled on every item it touches.

Good to know

Questions about Content Scorer

Usually under three seconds. We fetch your page once, count what can be counted and ask our decision model eleven questions in one go: six judgements about the page and five times 'what is the most important change here?'. Those last five are picked from a fixed list of forty changes, so you never get invented advice.

From check to system

Scoring one page is useful. Scoring every page is the system.

Traze writes content that already passes this check before it goes live, publishes it into your CMS and then measures per engine whether it gets cited. If a page falls behind, Traze rewrites it — you approve.

One check is a snapshot

Traze is the loop.

Traze researches, writes, publishes and optimises the content that gets you found in Google and in AI answers — automatically, with you in the loop. See pricing or book a demo.