What Is llms.txt and Do You Need One?
llms.txt is a plain-text file published at your domain root that gives AI systems a curated summary of your site. As of 2026 it's on roughly one in ten sites, and no major AI provider has committed to reading it in production — most AI search crawlers still read your HTML directly. It's worth adding because it costs nothing, but it isn't a substitute for real technical SEO and AEO work.
llms.txt is a proposed convention, first floated in 2024, for giving AI systems a curated, machine-readable summary of a site — a markdown-style plain-text file published at yourdomain.com/llms.txt, sometimes paired with a longer llms-full.txt. The pitch was straightforward: robots.txt tells crawlers what they can access, so why not have a file that tells AI systems what a site actually is, in language built for a model rather than a human skimming a homepage.
Here's the honest state of it in 2026: adoption is still low. Industry analysis of roughly 300,000 domains found about one in ten has an llms.txt file in place. More importantly, no major AI provider — not OpenAI, Google, Anthropic, or Meta — has publicly committed to reading or prioritizing it in production. The AI search crawlers that actually matter for citations, like GPTBot, ClaudeBot, and PerplexityBot, overwhelmingly still crawl and parse your live HTML directly rather than fetching the file.
Where llms.txt genuinely does get used today is a narrower case: developer tooling. Coding agents and IDE assistants — Cursor, Windsurf, Claude Code, GitHub Copilot, and similar tools — routinely check for /llms.txt when they're pointed at a documentation site, because it gives them a fast, reliable map of what's there. If you sell developer tools or maintain public documentation, that's a real, practical reason to have one, independent of anything to do with AI Overviews or ChatGPT citations.
For most businesses, the honest answer is: add it, but don't expect it to move your visibility. We publish one on this site — you can see it at /llms.txt — and it's generated automatically from the same content that drives the rest of the site, so it can never say something a live page contradicts. It costs nothing to maintain and there's no real downside to having it. What it isn't is a shortcut around the work that actually gets you cited.
That work is the same foundation AEO and GEO have always required: clean semantic HTML, fast Core Web Vitals, complete schema markup, and content structured so a model can extract a direct answer from it. Every major AI crawler already reads that directly off your pages. llms.txt is a nice-to-have on top of a well-built site, not a replacement for building one.
If you're deciding whether it's worth an afternoon of engineering time: yes, because the cost is close to zero and adoption could still shift. Just don't let a client or a vendor sell it to you as the thing that gets you into AI Overviews or ChatGPT answers — that's still earned through the technical and content work underneath it, not a text file sitting on top.