Content structure converts cleanly to markdown
A page whose HTML can be extracted to clean markdown is easier for models to consume.
Why the markdown lens
Most AI crawlers do not store your HTML verbatim. They convert it to markdown or plain text and keep the cleaned-up version. Pages that convert cleanly — with recognizable headings, lists, paragraphs, and code blocks — survive that conversion with their structure intact.
Pages full of nested divs, inline styles, and visual-only layout elements lose their structure in conversion. What was a clear hierarchy in the browser becomes a wall of undifferentiated text to the model.
How to be markdown-friendly
- — Use real HTML elements: h1-h6, p, ul, ol, blockquote, code, pre.
- — Avoid replacing headings with styled divs.
- — Do not build lists out of divs-with-bullets. Use ul or ol.
- — Wrap code samples in pre and code tags.
- — Skip decorative elements that have no text meaning.