Michaelliv/markit v0.5.3

on GitHub

Opportunistic markdown source discovery

markit now fetches raw markdown from docs sites instead of converting their HTML. Four detection methods, in priority order:

• Accept: text/markdown header on every request (Cloudflare, Vercel)
• <link rel="alternate" type="text/markdown"> tags in the HTML
• VitePress marker detection → fetches .md source files
• /llms.txt for root URLs (Stripe, Anthropic, etc.)

No extra requests for normal sites. The second fetch only fires when a markdown source is confirmed in the response.

Closes #13

Full Changelog: v0.5.2...v0.5.3