Back to Tools

Semantic Page Structure

Topical Authority, Heading Hierarchy & Keyword Relevance Studio
Topic: – 0 keywords 0 sections

Target Page & Overview

URL Fetch

For pages that block scrapers or render with JavaScript: open the page, copy its rendered HTML (DevTools → Elements → right-click <html> → Copy outerHTML) and paste it here. Keep the URL above filled in so the slug is scored too.

Fetching & analysing page structure…
Quick Samples
On-Page Relevance Model (0–100)

Extracts 1–4 word n-grams, merges singular/plural stems and scores each phrase across placement zones (Title 10, H1 8, URL 6, H2 5, Meta 5, H3 3, H4–H6 2, Alt/Bold 1.5, Anchor 1), body frequency (count0.7), opening-copy presence and section spread.

Ready On-page semantic & structural analysis

Structure & Topical Output

Map a page’s topical authority

Enter a URL on the left (or paste rendered HTML). The page is scraped, its heading hierarchy rebuilt and every keyword and key phrase scored for how strongly the page signals it to search engines.

You’ll get an interactive knowledge graph, a downloadable keyword and section table, a plain-text topical outline and a section-by-section alignment report.

Awaiting page analysis Exports: SVG · CSV · TXT · Markdown · PDF

About Semantic Page Structure

Semantic Page Structure scrapes a web page (or analyses pasted HTML), strips navigation, footers and boilerplate, rebuilds the heading hierarchy and scores every 1–4 word phrase (0–100) for how strongly the page signals it to search engines.

Output Views

ViewWhat it shows
Knowledge graphInteractive D3 force-directed network and collapsible tree visualising either Topic Clusters (main topic → sub-topic clusters → keywords) or Heading Hierarchy (H1 → H2 → H3 sections, targeted heading keywords and supporting copy terms) with optional sentence co-occurrence links and SVG export.
DataSortable and filterable 17-column keyword table with relevance (0–100), tier, cluster, frequency, density, placement flags (title, H1, URL, meta, H2, H3, bold, alt, anchor) and first-seen position, plus one-click CSV downloads for keywords and sections.
TextPlain-text topical outline of top keywords, topic clusters and the heading → keyword → supporting-terms tree, ready to copy or download as .txt.
ReportWritten diagnostic report with a main-topic placement checklist, top keywords, topic clusters, section-by-section alignment breakdown, on-page issues and Markdown/PDF export.

Questions

How is keyword relevance (0–100) calculated?

Every 1–4 word phrase that does not start or end with a stop word is extracted and stemmed so singular and plural forms merge. Each phrase is scored using placement weights (title 10, H1 8, URL slug 6, H2 5, meta description 5, H3 3, H4–H6 2, bold and image alt 1.5, anchor text 1), body copy frequency with diminishing returns (count^0.7), an opening-copy bonus (+2 in the first 10% of words), an exact title/H1 match bonus (+20), a specificity multiplier for multi-word phrases and a section-spread multiplier. Fragments that appear almost exclusively inside a longer phrase are discounted or removed, and scores are normalised to 0–100.

How are topic clusters built?

Keywords are grouped by their strongest shared stemmed term (weighted by relevance). Terms belonging to the page’s main topic are reserved as a last resort so the clusters reveal distinct sub-topics rather than collapsing into a single group.

What does the section alignment score measure?

Section support (0–100) blends three on-page signals: whether the heading’s target keywords recur in that section’s copy (50%, with partial credit when individual words appear), how many supporting copy terms are semantically related to the heading (30%) and section depth (20%, reaching full credit at 150+ words).

What if a page blocks scrapers or renders with client-side JavaScript?

Click “Paste HTML instead” in the left panel, copy the rendered outerHTML from your browser’s DevTools Elements panel and paste it into the studio. Keep the page URL filled in so the URL slug is still included in the scoring.