Group pages by what they say
Cluster by embedding so pages sit together because they mean the same thing, not because they share a folder.
Link the pages that belong together.
A link graph from meaning, not from the sitemap.
Most internal linking work starts from the links you already have. It cannot tell you that two pages are a few degrees apart in meaning and have no path between them — usually the cheapest win on the site.
How the graph gets built
Find closeness with no path
Filter for pages that are semantically close and unconnected — the missing edges this site actually needs.
Ship the pairs, not a theory
Export source and target URLs as a sheet, with the passages that justified each link still attached.
The work it takes off your plate
Embeddings for every page
Represent page content as searchable meaning, even when URLs do not share the same keywords.
Read the link graph against meaning
Compare what links to what with how closely the two pages actually relate.
Semantic linking gaps
Find pages that are close in meaning but poorly connected, then turn those pairs into work.
Intent overlap detection
Spot clusters where several URLs compete for one answer and decide what to merge or differentiate.
Existing graph overlay
Lay current internal links on the map so the missing edges are specific, not a generic “link more.”
Topic coverage maps
See where a subject is well supported, where coverage is thin, and which groups should carry authority.
Briefs grounded in URLs
Write linking recommendations that point back to the pages and passages that justify them.
Catch thin and redundant pages
Low-substance URLs and two pages competing for one intent, flagged by similarity.
Entity and concept review
See how brands, products, and repeated concepts show up across templates and clusters.
Related use cases
Give agents SEO evidence, not slop
Crawl, chunk, and embed any site on your machine, then hand your agents the passages and page evidence behind every answer.
free to start