Indexing a whole site

A single article becomes a study space well. A documentation set, a course outline or a multi-part guide does not: the useful material is spread across pages that link to each other. The site tool maps those links, ranks the pages that carry real content, and builds a library from the set rather than from one entry point.

• Pages with substantial readable text, rather than navigation shells.
• Pages inside the same site, so an index does not wander off across the web.
• A bounded number of pages, so one request cannot crawl an entire domain.

Index pages that exist only to link elsewhere, duplicate print views, and anything the site's own rules ask automated readers not to fetch. Dropping them is deliberate: a study library of twelve navigation pages is worse than a library of four real ones.

Each kept page becomes its own local project with its own claims and evidence, and they appear together in My Library. You can study one, export it, or delete the whole set. Nothing about the index is stored anywhere but this browser.