Serving Markdown to LLMs with Eleventy

A walkthrough on generating Markdown twins of every blog post using Eleventy, so AI crawlers and LLMs can consume content more efficiently. Serving Markdown instead of HTML reduces token usage by 68–97%, saving cost and context window space. The implementation uses Eleventy's pagination feature to create a second layout (post-markdown.liquid) that outputs YAML front matter plus raw Markdown content at a parallel URL (index.md). Additional steps cover advertising the Markdown files via rel="alternate" link tags, deindexing them from search engines using X-Robots-Tag headers on Netlify, and disabling generation in development to avoid slowdowns. Content negotiation via Accept: text/markdown is discussed but dismissed since no AI crawler currently uses it.