What llms.txt is — and an honest note on adoption
llms.txtis a proposed standard for a markdown file at your domain root that gives AI systems a curated map of your site: an H1 with the name, a blockquote summary, and sections of annotated links. The pitch is reasonable — language models have small context windows and your site has hundreds of pages, so a human-curated index helps them find the canonical answer instead of scraping whatever ranks. But let's be honest about where adoption stands: a 2026 Ahrefs study found most llms.txt files are never fetched by AI bots, and none of the major AI providers have formally committed to reading it. If someone sells you llms.txt as an AI-traffic growth hack, they are selling vapor.
So why bother?
Because the cost is fifteen minutes and the downside is zero. That is the entire hedge math. The file is a few kilobytes of static text — no build step, no maintenance burden beyond updating links when your site restructures, no risk of hurting rankings (Google ignores it entirely). If the convention wins, early adopters have a well-curated file that agent frameworks and AI crawlers already understand. If it dies, you have lost nothing and gained a decent one-page summary of your own site, which is a useful editorial exercise in itself. This is the same posture we take in the full RankVitals audit: we check for llms.txt and note it as a low-severity, low-effort item — never as a critical issue.
How to write a good one
Keep it short and opinionated. The site name goes in the H1, and the blockquote should say what the site is and who it serves in one or two sentences — write it like the answer you wish an AI assistant would give when asked about you. Then list your key pages: product overview, documentation, pricing, your best guides. The per-link notes matter most, since they tell a model which page answers which question ("API reference with code examples", "pricing including the free tier"). Skip legal boilerplate, tag archives and anything you would not show a new customer. Generate the file above, save it as llms.txt, deploy it at the root, and verify it loads as plain text at https://your-site.com/llms.txt.