https://www.mohit.dev stays HTML for people. Agents can ask for the same page as Markdown.
Accept: text/markdown on a post returns the built .md file, with a canonical link back to the HTML. Browsers still get HTML. The same files are also at the .md URL.
/llms.txt lists the site, the RSS feed, and recent posts. robots.txt says ai-train=no, search=yes, ai-input=yes. The homepage links to /llms.txt and /rss.xml.
No API, so no API catalog and no DNS agent records. A static site plus one Cloudflare Worker is enough.