Starting with AI-assisted internal links

I have started building a WordPress plugin to make internal linking less manual. The first version centres on keyword rules: each rule associates a phrase with a target post, and the plugin can use those rules when adding links to content.

From keywords to suggestions

In the post editor I can manage the rules associated with a page, including the keyword, priority, link limit and post types. An AI-assisted action reads the page title and content and suggests specific phrases that could link back to it. The prompt asks for phrases that describe the destination accurately rather than generic words that could lead readers somewhere unexpected.

There is also an opportunity finder for the post being edited. It supplies the model with available target keywords and asks for an original sentence, a minimally changed sentence containing the keyword, and a reason. The interface shows the proposed change alongside the original, with a paragraph indicator to help locate it. I can select suggestions before applying them.

Managing the links

Applying a suggestion replaces the matching sentence in the stored post content. The implementation tries an exact match and a decoded-text fallback; if it cannot find the original, it reports that rather than claiming a change was made. The rule editor and link replacement controls also received several early refinements.

I added a separate interlink management dashboard to provide a view beyond one post at a time. This is an early rule-based workflow, with model suggestions helping prepare and adapt the text. It is a starting point for experimenting with how much of the repetitive work WordPress can handle.