In September 2024, fast.ai founder Jeremy Howard proposed a new standard called llms.txt. The idea: give AI language models a clean, structured summary of your website — the same way robots.txt gives web crawlers access rules and sitemap.xml gives search engines a list of your pages.
A year later, several AI platforms and tools have started checking for llms.txt, and it's become one of the signals GEO audits measure. Here's what it is and how to create one.
What problem does llms.txt solve?
When an AI model visits your website, it has to parse HTML, ignore navigation, headers, footers, ads, and cookie banners, and extract the actual content. Most web pages weren't designed to be read by AI — they were designed to be rendered by browsers.
llms.txt is a plain-text file at the root of your domain that provides AI models with a clean, prioritized summary of your site: what it is, who runs it, and which pages matter most. No HTML to parse, no noise to filter.
What goes in an llms.txt file?
The format uses Markdown with a specific structure. A minimal but complete llms.txt looks like this:
# GEO Auditor
> AI search visibility auditor that scores your website across 40+ signals
> and delivers a prioritized fix plan for ChatGPT, Perplexity, Gemini, and Claude.
GEO Auditor is a SaaS tool for measuring and improving AI search visibility
(Generative Engine Optimization). It checks AI crawler access, schema markup,
brand entity recognition, and content citability.
## Docs
- [How it works](/about): Overview of the audit methodology and signals checked
- [Pricing](/): Free tier with GEO score; $29 one-time for full report with fix plan
## Blog
- [What is GEO?](/blog/what-is-generative-engine-optimization)
- [Fix GPTBot in robots.txt](/blog/gptbot-robots-txt)
- [How to get cited by ChatGPT](/blog/how-to-appear-in-chatgpt-answers)The key components:
- H1 heading: Your site or organization name
- Blockquote: One or two sentences describing what the site does — this is the most important part, as AI models often use it as a summary
- Body paragraph: A few sentences of additional context
- Sections with links: Categorized links to your most important pages, each with a brief description
How to add llms.txt to your site
Create a plain text file named llms.txt and place it in your site's public root, so it's accessible at https://yourdomain.com/llms.txt.
Next.js / App Router
Put the file in your public/ directory:
public/llms.txtIt will be served statically at the root URL with no additional configuration.
Static sites (Hugo, Jekyll, Eleventy)
Put llms.txt in your site's static root. Most static site generators serve files from a static/ or public/ folder at the root URL.
Vercel / Netlify
Same as above — any file in public/ is served at the root. No configuration needed.
Should you also create llms-full.txt?
The spec defines an optional companion file, llms-full.txt, which contains the complete text of your site's key pages — without any HTML. This is useful for AI tools that want the full content of your documentation or blog in a single, parseable file.
For most sites, start with llms.txt alone. Add llms-full.txt if you have documentation, tutorials, or reference content that you want AI tools to be able to read comprehensively.
How much does llms.txt matter for GEO?
Directly: it's a moderate signal, not a major one. No AI platform has confirmed that the presence of llms.txt directly increases citation frequency.
Indirectly: it's a strong signal of AI readiness. Sites that have an llms.txt file tend to have also addressed the other GEO fundamentals — crawler access, schema markup, and content structure. It's a 10-minute task that fills a visible gap in your AI visibility profile.
To check whether your site has llms.txt and see how it stacks up against the rest of your GEO signals, run a free audit.