Internal link SEO plugin — vector-similarity, not string match.
Most internal-linking plugins keyword-match a target phrase against post titles. We build a vector-embedding index of every post and suggest links by semantic similarity — the same technique Google uses. Result: relevance jumps, over-linking drops.
Why string-match plugins under-perform
String matchers over-link common phrases and miss semantically related pages. A post about 'schema markup' never gets linked to a post about 'JSON-LD for events' because the words don't overlap — even though a human would link them instantly.
How vector similarity works here
We embed every post with a lightweight local model, index in a vector store per site, and suggest links at edit time with a similarity score + a diminishing-returns limit per anchor. Cost per site: ~$0.02 / month at 500 posts.
What you see in wp-admin
Sidebar widget with 8 suggested links per post, each with target URL, suggested anchor and similarity score. One click inserts, one click dismisses (and dismissal is remembered).
Common questions
- Does this replace an SEO editor?
- No. It removes the mechanical work so the editor spends time on the calls it should be making — narrative, angle, headline.