Live demo
See it work.
Replay a real crawl, read the two files it writes, then compare one page before and after extraction.
Output
Watch it run, then read the files.
Press run to replay a real crawl of the tool's test site: a small company site with docs, a blog, a robots.txt, a gzipped sitemap index, a noindex page, a duplicate and a 404.
The finished files
Before and after
Only the content survives.
The same page as served and as extracted. Navigation, sidebar, cookie banner, footer, script and the heading permalink are removed. The code keeps its language, the table stays a table, and links become absolute.
HTML as served
Markdown extracted