llms.txt is a proposed Markdown file at your site root that gives language models a curated, machine-readable map of your most important content. It’s an emerging convention — not yet a universal standard — so here’s an honest look at what it does, what it doesn’t, and how to write one.
The convention is simple Markdown: an H1 with your site name, a short blockquote summary, then H2 sections grouping links to your most important pages with one-line descriptions.
robots.txt grants or denies crawl access (allow GPTBot, PerplexityBot, ClaudeBot and Google-Extended there). sitemap.xml lists every URL for completeness. llms.txt is the editorial layer — a short, prioritised, human-readable guide to what actually matters. Use all three; they don’t compete.
Honestly: it’s low-effort and low-risk, but adoption across engines is still uneven, so don’t expect dramatic results on its own. Prioritise the fundamentals first — unblock AI crawlers, fix the render gap, add structured data — then add llms.txt as a finishing touch.
A proposed Markdown file at your site root (/llms.txt) that gives AI models a curated, prioritised map of your most important content and how to read it.
No. robots.txt controls crawl permissions; llms.txt is an editorial guide to your best content. They’re complementary.
Support is emerging and varies by engine — it isn’t a guaranteed, universal standard yet. It’s a cheap, sensible addition, but fix crawlability, rendering and schema first.
Run Beacon’s free AI Visibility check — it reports whether /llms.txt exists alongside your AI-crawler access and on-page signals.
See your llms.txt status, crawler access and schema in seconds.