What llms.txt is supposed to do
llms.txt is a proposed markdown file, sitting at the root of a domain, meant to give AI systems a clean, curated summary of a site’s content — the pitch is that it works like robots.txt or a sitemap, but written for language models instead of crawlers.
The idea gained traction quickly, with vendors marketing it as an essential step for “AI SEO readiness” — but adoption and actual evidence of impact have told a very different story from the pitch.
What Google has actually said about it
Google’s own documentation states plainly that a site doesn’t need new machine-readable files, AI text files, markup, or Markdown to appear in Google Search or its generative AI features — Google Search doesn’t use them. That’s about as direct a denial as a platform issues.
This matters specifically because Google AI Overviews are one of the highest-visibility AI surfaces most businesses care about, and Google is telling site owners directly that llms.txt has no bearing on it.
What the broader research shows
Beyond Google’s own statement, independent testing found llms.txt made no measurable difference to whether a domain got cited by AI models generally — in one study, the tested model actually performed slightly better without it. Adoption reflects the same skepticism: scans of the most-visited sites globally find well under 1% actually publish one.
None of this means the underlying goal — being legible and citable to AI systems — is wrong. It means this specific file isn’t how that goal gets achieved; the fundamentals covered in our GEO guide are what actually move that needle.
Where it might still be worth the effort
The clearest legitimate use case is developer tooling: AI coding assistants increasingly parse documentation to help developers integrate an API or SDK, and a clean, curated llms.txt can genuinely make that parsing more reliable. Named early adopters — Anthropic, Stripe, Cloudflare, Vercel — are almost entirely developer infrastructure companies, not general content sites.
If your business isn’t a SaaS product, API, or documentation-heavy platform, this narrow benefit largely doesn’t apply, and the time is better spent elsewhere.
What actually drives AI visibility instead
Server-rendered, crawlable HTML, direct and well-structured answers, and genuine authority signals are what AI systems actually rely on when deciding what to cite — the same fundamentals covered in detail here. None of that is replaced or shortcut by adding a text file Google has already said it ignores.
Treat llms.txt as a low-cost, low-priority addition for the narrow audience it actually serves — not a required item on a general AI-SEO checklist.