What it is.
You hand it a live page URL or paste raw markdown. It scrapes the page if needed, crawls your sitemap to find up to 500 site URLs as candidate targets, groups those candidates into tiers, then writes links into the text and hands back markdown.
Product pages, feature pages, blog listicles, other blogs, and everything else sit in separate priority bands, so the commercially valuable destinations win ties.
Slate's own agent, so running it needs a Slate account. The four steps and the tier list are published, which is enough to rebuild the same logic elsewhere.
What you get.
- A link target list built from your own sitemap, capped at 500 URLs.
- Page tiers that favor product and feature pages over another blog post.
- Updated content returned as markdown, with existing links preserved.
- Natural anchor text chosen per insertion instead of a repeated exact-match phrase.
How to set it up.
Decide your tier order before anything else, since that single decision drives every link it picks.
Point it at a sitemap that matches your live site, because a stale sitemap produces links to dead pages.
Start with one published page before you try a batch, and read every inserted link.
Check that existing links survived, since that's what breaks when a tool rewrites a draft.
Set a link ceiling per piece yourself, because the agent will happily find more targets than a reader wants.
Re-run it over old posts once you trust the tier order, oldest and highest-traffic first.
Use cases
Build a target list from your sitemap
Generate the capped list before anyone writes an anchor.
Favor the pages that convert
Use the tiers so product pages get the links ahead of blog posts.
Rebuild an archive
Run it at site scale instead of repairing one post at a time.
Best for
A blog with no link structure
Hundreds of posts, nobody linking between them, and no appetite for a manual audit.
Pre-publish link passes
A new draft has no inbound context, so the links have to be added before it goes live.
Copying the logic
The tier list and the 500-URL cap are the design decisions; you can rebuild both in any tool.