← Back to Blog

What Is llms.txt?

Your website has a robots.txt that tells search engines what to crawl. It has a sitemap.xml that lists every page. But when ChatGPT or Perplexity tries to understand your business, neither of those files helps much.

That's where llms.txt comes in — a simple file that gives AI models a clean, human-readable overview of your website. It's the newest standard in web visibility, and most businesses don't have one yet.

This article is part of our AI Visibility Glossary — plain-language definitions of every term you need to understand how AI models find, evaluate, and recommend businesses online.

What Is llms.txt?

llms.txt is a plain-text file you place at the root of your website (at /llms.txt) that provides a concise, structured summary of your site's content — written specifically for large language models (LLMs).

It was proposed by Jeremy Howard, co-founder of Answer.AI, in September 2024. The idea is simple: give AI models a cheat sheet about your website so they don't have to crawl and interpret messy HTML with navigation menus, ads, and JavaScript.

Think of it this way: robots.txt tells bots what they can access. llms.txt tells AI models what they should read.

Why Does llms.txt Matter?

Large language models have a fundamental limitation: their context windows are too small to process an entire website. When ChatGPT or Perplexity wants to learn about your business, it can't just load every page you have — it needs a curated summary.

Without llms.txt, AI models have to:

With llms.txt, you hand them a clean Markdown file that says: "Here's who we are, here's what matters, and here are the detailed pages you should read."

2.5K+GitHub stars on the llms-txt repo
90%Of pages crawled by AI bots published within 3 years
0Cost to implement llms.txt

How Does llms.txt Work?

The file uses Markdown — a format that both humans and AI models can read easily. It follows a specific structure:

# Your Business Name

> A short description of what your business does and who you serve.

## Products

- [MenuBoost](https://hd-webdesign.si/menu-boost/) : AI-powered menu descriptions for restaurants
- [BoostSuite](https://hd-webdesign.si/boost-suite/) : Marketing toolkit for agencies

## Resources

- [About Us](https://hd-webdesign.si/about/) : Company story and team
- [Pricing](https://hd-webdesign.si/pricing/) : Plans and pricing details

The structure is simple:

  1. H1 heading — Your business or project name (required)
  2. Blockquote — A short summary of what you do
  3. H2 sections — Lists of important pages grouped by topic
  4. Links — Each entry is a Markdown link with an optional description

That's it. No complex schema, no special syntax — just clean Markdown that AI models already understand perfectly.

llms.txt vs. robots.txt vs. sitemap.xml

These three files serve different purposes, and a well-structured website has all three:

File Purpose Audience Format
robots.txt Tells bots which pages they can or cannot crawl Search engine crawlers Plain text
sitemap.xml Lists all pages on your site for indexing Search engines XML
llms.txt Provides a curated summary for AI understanding AI models (ChatGPT, Claude, Gemini) Markdown

The key difference: robots.txt and sitemap.xml are designed for traditional search engines. llms.txt is designed for AI systems that generate answers rather than show lists of links.

They coexist peacefully. llms.txt doesn't replace either of the other files — it adds a new layer specifically for AI visibility.

Who Should Create an llms.txt File?

Every business that wants to be visible to AI search should have one. But it's especially valuable for:

The adoption is growing fast. Projects using nbdev now create Markdown versions of all pages by default, and the FastHTML project was one of the first to implement the full llms.txt specification.

How to Create Your llms.txt File

Step 1: Identify Your Key Pages

List the 10-20 pages that best represent your business. Think about what an AI model would need to know to recommend you accurately:

Step 2: Write the Markdown

Create a file called llms.txt following the format above. Keep descriptions short — one sentence per link is enough. Group related pages under H2 headings.

Step 3: Place It at Your Root

Upload the file to https://yourdomain.com/llms.txt. It must be accessible at the root path, just like robots.txt.

Step 4 (Optional): Create Full-Content Versions

The llms.txt spec also suggests providing clean Markdown versions of key pages at URLs with .md appended. For example, your pricing page at /pricing.html could also be available at /pricing.html.md. This gives AI models access to full page content in a clean format.

Want to check if your site already has an llms.txt file?
Visit https://yourdomain.com/llms.txt in your browser. If you see a 404, you don't have one yet. Run a free AI Visibility Audit with BoostSuite to see how AI models currently see your website.

Real-World Example: HD WebDesign

Here's what a practical llms.txt looks like for a web design business:

# HD WebDesign — AI-Powered Web Design & Marketing

> Web design, AI chatbots, and marketing tools for businesses
> in Slovenia, Croatia, Italy, and the DACH region. Products:
> MenuBoost (restaurant menus), BoostSuite (agency toolkit).

## Services

- [Web Design](https://hd-webdesign.si/web-design/) : Custom websites from €890
- [AI Chatbots](https://hd-webdesign.si/boost-suite/) : AI-powered customer support

## Products

- [MenuBoost](https://hd-webdesign.si/menu-boost/) : AI menu descriptions, 10 languages, €19/mo
- [BoostSuite](https://hd-webdesign.si/boost-suite/) : SEO audit, GEO check, ad copy, listing optimizer

## Resources

- [Blog](https://hd-webdesign.si/blog/) : Articles on AI, GEO, and web design
- [About](https://hd-webdesign.si/about/) : Team and company information

This gives an AI model everything it needs to accurately describe and recommend the business — in under 200 words.

What Happens If You Don't Have One?

AI models will still try to learn about your website — they'll crawl your pages, parse your HTML, and guess what's important. But the results are unpredictable:

As AI search grows — ChatGPT alone now has 900+ million weekly users — the businesses that make it easy for AI to understand them will have a measurable advantage.

FAQ

Is llms.txt an official standard?

It's an open proposal by Jeremy Howard (Answer.AI), published in September 2024. While not an official W3C standard, it's gaining rapid adoption. The GitHub repository has over 2,500 stars, and major projects like FastHTML and nbdev already implement it. Think of it like how robots.txt started — as a practical convention that became universal.

Will llms.txt replace my sitemap?

No. llms.txt complements your sitemap. The sitemap lists all your pages for search engine indexing. llms.txt provides a curated summary for AI understanding. A well-optimized website has all three: robots.txt, sitemap.xml, and llms.txt.

Do I need technical skills to create one?

Not really. If you can write a list in plain text, you can create an llms.txt file. It's just Markdown — a heading, a description, and a list of links. No coding required. If your website already has a robots.txt, you can create an llms.txt in the same way.

How often should I update my llms.txt?

Update it whenever you add new products, services, or important pages. For most businesses, once a quarter is enough. The file should always reflect your current offerings and most valuable content.

Do AI models actually read llms.txt files?

Yes. AI systems that browse the web — like ChatGPT with browsing, Perplexity, and Claude — can access and parse llms.txt files. The Markdown format is already native to how LLMs process text, making it one of the most efficient ways to communicate with AI models.

This article is part of the AI Visibility Glossary — plain-language definitions of every term you need to understand how AI models find, evaluate, and recommend businesses online.

Related articles: