llms.txt

llms.txt is a plain-text Markdown file published at the root of a domain (e.g. example.com/llms.txt) that offers large language models a concise, curated index of a site's key pages — often with links to clean Markdown versions of them — so an AI can find and ingest the most relevant content without crawling and parsing full HTML. Proposed in 2024 (llmstxt.org), it's modeled in spirit on robots.txt and sitemap.xml, but its purpose is discovery and curation, not access control.

That distinction matters: llms.txt doesn't block or permit crawlers — it helps cooperating AI tools spend their limited context window on your best, cleanest content instead of guessing at it. Sites often pair it with an expanded llms-full.txt that inlines the full text of the catalog. Adoption is early and voluntary; no major model provider guarantees it reads the file, but it's becoming a common signal in AI-era discovery.

For SEO and GEO, llms.txt is about making your site legible to answer engines and AI assistants. If you want to be cited, handing models a clean, structured map of your authoritative pages lowers the friction for them to retrieve and attribute you. Treat it as complementary to structured data and a clean sitemap — not a replacement for either, and not a way to keep AI out (that's robots.txt and emerging content-signal standards).