using this as a preprocessing step before feeding pages to an LLM, way less noisy than piping raw HTML through and stripping tags myself. the markdown output stays clean even on pages with a lot of nested divs and ad clutter, which is where most of the DIY approaches I tried before this fell apart