llms.txt

llms.txt is a proposed convention, not a formal web standard, where a site publishes a plain-text file at its root listing key pages, summaries, and links intended to help AI systems understand its content quickly. It is modeled loosely on robots.txt but has no enforcement mechanism and no guarantee that any given crawler or model actually reads it.

The file typically lives at /llms.txt, uses simple markdown-style headers and links, and is meant to give an AI system a curated, low-noise entry point to a site’s most important content instead of forcing it to crawl and parse full HTML pages. No major AI company has publicly confirmed that its systems read the file at all.

Independent testing to date has found no measured citation lift from adding an llms.txt file, and adoption sits at roughly five to ten percent of top sites. Publishing one is best understood as a forward-compatibility bet rather than a proven visibility lever: low cost to add, unproven return, worth revisiting as adoption and crawler support evolve rather than treated as a solved tactic today.