llms.txt is a plain markdown file placed at the root of your website (yoursite.com/llms.txt) that gives AI systems a curated map of your most important content. It was proposed by Jeremy Howard of Answer.AI in September 2024 as a way to help large language models find and read the right pages on your site without wading through navigation, scripts, and clutter.
That is the simple definition. The harder, more useful question is whether AI engines actually use it, and whether adding one will do anything for your brand. This guide gives you the honest answer, a working example you can copy, and a way to check what your own site is doing.
What llms.txt is designed to do
A normal web page is built for browsers and humans: HTML, menus, banners, scripts, and the actual content buried somewhere in the middle. That is noisy for a language model trying to understand what your site is about.
llms.txt is meant to fix that. It is a single markdown file with a clear structure: your brand name, a one-line description, an optional paragraph of context, and a curated list of links to your most important pages, each with a short description. The idea is that an AI system can read this one file and immediately understand what you offer and where the good content lives, instead of guessing.
There is also a companion file, llms-full.txt, which contains your full site content in one markdown document for deeper ingestion. Most sites only need the basic llms.txt. Documentation-heavy products sometimes maintain both.
How to create one
Creating a basic llms.txt takes under 30 minutes. The format is just markdown. Here is a working example for a real site, structured the way the spec recommends:
# Rankry
> Rankry is an AI visibility platform that tracks how brands are
> cited and recommended across ChatGPT, Claude, Gemini, Perplexity,
> and Grok, then gives them a plan to improve.
## Product
- [Features](https://rankry.ai/features): Tracking, audits, and the action-plan tools Rankry offers
- [AI Readiness](https://rankry.ai/features/ai-readiness/): How Rankry audits a site's technical readiness for AI engines
- [Pricing](https://rankry.ai/pricing): Plans and what each one includes
## Guides
- [The 5-Layer AI Visibility Stack](https://rankry.ai/blog/): The framework brands move through to get recommended by AI
- [Beyond Mention Monitoring](https://rankry.ai/blog/): Why position and reasoning matter more than a yes or no mentionThe rules are short. Start with an H1 containing your brand or site name. Add a blockquote with a one-line summary. Use H2 sections to group links. Each link is a markdown link followed by a colon and a short, plain description. Keep it to your genuinely important pages, not every URL on the site.
You do not have to write it by hand. On WordPress, Yoast SEO and Rank Math both generate the file automatically. Webflow lets you upload it to your root directly. Next.js, Astro, and Mintlify have plugins or built-in support. Once the file exists, publish it at yoursite.com/llms.txt and you are done.
Do AI engines actually use it?
This is where most articles either oversell or dodge. Here is the honest read as of mid-2026.
No major AI provider has publicly confirmed that it uses llms.txt to rank or cite content in its answer surfaces. That includes OpenAI, Google, Anthropic, Meta, and Mistral. Google has been the most direct: its team confirmed in 2025 that it does not support llms.txt and is not planning to, and its May 2026 AI optimization guide explicitly lists llms.txt as something that is not needed for AI Overviews, AI Mode, or any other generative search feature.
The independent evidence points the same way. SE Ranking analyzed close to 300,000 domains and found no correlation between having an llms.txt file and being cited by AI systems. Semrush ran a controlled study and found no statistical correlation with improved AI performance. One analysis of over 500 million AI bot visits across 90 days found only a few hundred requests for llms.txt directly. Adoption itself sits around 10% of sites and is not growing quickly.
So where does it genuinely get used? Developer tooling and agents. AI coding assistants like Cursor, GitHub Copilot, and Claude Code read llms.txt to pull documentation efficiently with less wasted context. That is why the companies shipping it most seriously are developer platforms: OpenAI and Anthropic both publish llms.txt for their own docs, Anthropic recommends it in its guidance for agent-facing content, and Stripe, Cloudflare, and Vercel all maintain one. Chrome’s Lighthouse added an audit for the file in May 2026. There are also early reports that some answer engines may read it to help prioritize pages, but none of that is broadly confirmed.
The honest bottom line: llms.txt is cheap, low-risk infrastructure. If you run developer documentation or an API reference, or you simply want to be forward-compatible when platforms formalize support, add one. It will not hurt you. But it is not a citation lever today. It will not move your visibility in AI answers on its own, and treating it as a shortcut is the mistake. The work that actually drives AI citation is entity-consistent content, direct-answer page structure, third-party presence on sites AI trusts, and tracking your citations across the major engines.
Common mistakes
The most common and most damaging mistake is generating a separate markdown copy of every page on your site and linking them all from llms.txt. If those markdown files are indexable, you have just created large-scale duplicate content, which dilutes crawl budget and can suppress the rankings of your original pages. Keep llms.txt to a curated list of your genuinely important pages.
A few others worth avoiding: stuffing the file with every URL instead of your best ones, writing vague descriptions that tell a model nothing, letting it go stale after a site redesign, and treating llms.txt as a replacement for the real AEO work rather than a small piece of it.
How to check yours
Once your file is live, you want to confirm it exists, parses correctly, and is reachable by AI crawlers, and then make sure you are spending your real effort on the things that actually move AI visibility.
Rankry’s AI Readiness Checker audits this directly. It detects whether your site has an llms.txt file, flags it if it is missing or malformed, and places that result inside the bigger picture: crawlability, structured data, entity recognition, and the other technical signals that determine whether AI engines can read and trust your site. Running rankry.ai through the checker, for example, shows the llms.txt status alongside the rest of the readiness profile, so you can see at a glance what is in place and what is not.
The point is to keep llms.txt in proportion. It is one quick checkbox in a much larger audit. The Checker is part of a full AI SEO site audit, and the technical readiness layer it covers is detailed on the AI Readiness page.
llms.txt is worth adding. It is not worth obsessing over. The brands that win in AI answers are not the ones with the prettiest text file, they are the ones tracking where they actually stand across every engine and fixing the gaps that matter.