What the page says it is about, against what it is about
Every page declares a topic in its title, headings, and schema, and demonstrates a different one in its body text. This measures both and reports the gap, including the terms a page genuinely covers but never claims credit for.
Declared vs Computed Topic
What the page declares it is about
What the body text is actually about, by TF-IDF
Term weights are real TF-IDF. Document frequency comes from a background corpus of pages fetched from this same origin, so a term like schema scores low here precisely because it appears everywhere on this site, while a term specific to this page scores high. That is the correct behaviour: salience is relative to the corpus, and a word that describes your whole site cannot also be what distinguishes one page within it.
Two topics, one page
Every page has a topic it claims in its title, heading, and schema, and a topic its body text actually supports. When those disagree, the page competes for neither well.
The gap is usually the title's fault
Titles get written for a keyword the body never earns. The body is the harder thing to change, which is why the mismatch survives so many content reviews.
Orphan terms are opportunities
A term with high salience that appears nowhere in the title, heading, or schema is something this page genuinely covers and gets no credit for. Those are the cheapest wins on any site.
Site wide terms cannot differentiate
If your brand's core noun appears on all 400 pages, it has near zero inverse document frequency and does nothing to distinguish any single page. Optimising harder for it does not help.