llms.txt is a plain-text file you put at the root of your site — at /llms.txt — that hands AI models a clean, curated map of your most important content. Think of it as a sitemap written for language models instead of search crawlers.
Is it a magic ranking trick? No. As of 2026 the major AI search engines have not confirmed using it to rank or cite pages, and tracking studies have not found a measurable traffic lift. But it is cheap to add, it is genuinely useful for AI coding agents and assistants, and it signals that you are thinking about machine readers. Do it for the right reasons — not because someone promised you a citation boost.
It is Markdown. A top-level heading with your name, a one-line summary in a blockquote, then sections of links with short descriptions pointing at the pages you most want an AI to read:
# Acme Inc > Acme makes payment APIs for developers. ## Docs - [Quickstart](https://acme.com/docs/quickstart): get your first payment in five minutes - [API reference](https://acme.com/docs/api): every endpoint, with examples ## Company - [Pricing](https://acme.com/pricing): plans and limits - [About](https://acme.com/about): who we are
Keep it short and high-signal. Link your docs, your key product and pricing pages, and anything you would want quoted accurately. Skip the marketing fluff — the goal is comprehension, not persuasion.
Here is where most articles oversell. The honest evidence in 2026:
For AI search citations (ChatGPT, Perplexity, Google AI): unproven. None of the major engines has confirmed reading llms.txt to decide what to cite, and independent tracking has shown no reliable lift. Treat any "llms.txt boosted my AI traffic" claim with skepticism.
For AI coding agents and assistants: genuinely useful. Developer tools like Cursor, Claude Code and GitHub Copilot routinely fetch /llms.txt when pointed at a documentation site, so a good file makes your docs easier for agents to use correctly.
Bottom line: add it if you have docs or want to be agent-friendly. It is low-cost and forward-looking. Just do not let it distract you from the signals that demonstrably decide whether an AI can read you at all — crawler access, server-rendered content, and structured data.
Web-wide, adoption is still tiny — roughly one in ten sites has an llms.txt. But among sites that take machine-readability seriously, it is mainstream: across the 154 leading sites Oraql audited for our 2026 State of AI Search Readiness report, 49% publish an llms.txt — nearly five times the web-wide rate. Early adopters skew technical (Cloudflare, Vercel and similar shipped one long ago).
1. Write the Markdown above, tailored to your site. List your five to fifteen most important pages.
2. Save it as a file named llms.txt.
3. Serve it at your site root so it resolves at yoursite.com/llms.txt (same place as robots.txt).
4. Keep it current when your important pages change.
That is it. No meta tags, no schema, no build step required.
Use our free llms.txt generator → — fill in your pages and copy a valid file.
Oraql checks whether your site has an llms.txt — and, more importantly, scores the seven signals that actually decide whether ChatGPT, Perplexity, Claude and Google AI can read and recommend you. It is free and takes seconds.