What is llms.txt, and does your website need one?

An llms.txt file is a plain Markdown file at the root of your website that summarizes your business and links to your most important pages, written for AI models and agents rather than people. Most business websites don't need one yet: Google says you don't need it for its search or AI features, and no major AI search tool has said it relies on it. It is most useful for software companies with documentation that developers read through AI coding tools, and it takes under an hour to make, so it costs little to add.

The idea comes from Jeremy Howard of Answer.AI, who published the proposal in 2024. Web pages are built for browsers, with navigation, scripts and layout code around the actual content. An llms.txt file gives an AI agent a short, clean map instead.

Why llms.txt was created

To understand the purpose of llms.txt, consider how a large language model reads an ordinary web page.

When an AI agent visits your homepage, it typically downloads the full HTML document. That file is packed with:

• Header navigation menus and mobile drawer code.
• Tracking pixels, cookie consent banners, and tag managers.
• Inline CSS styles, SVG icons, and interactive widget scripts.
• Footer disclaimers, copyright notices, and repeated link lists.

The code behind a page is often many times longer than the text a visitor reads. When an AI model reads the raw page, that code takes up space in its limited context and costs processing time, and the core facts (what your product does, what it costs) get buried.

The llms.txt proposal addresses this problem by offering a lightweight alternative. It lives at yourdomain.com/llms.txt and uses plain Markdown. An AI tool that supports the format can fetch this file and get a short summary plus a list of links to your most useful pages.

How the llms.txt file is structured

The proposal defines a simple Markdown layout, in this order:

1. H1 title: The name of your project, product, or organization. This is the only required part.
2. Summary blockquote: A two-to-three-sentence summary explaining what your company does and who it serves.
3. Details: Optional paragraphs or lists (no headings) with context for reading the file.
4. Link lists: H2 sections that group your main pages, each link written as - [Page title](URL): short description. A section titled "Optional" holds links an agent can skip.

Here is a practical example of a clean llms.txt file for a service business:

# Techcelerant

> Techcelerant is a digital agency that builds websites, online stores and apps and runs SEO, paid ads and other marketing for businesses across the United States.

All packages are fixed-price, with the payment schedule listed for each.

## Services

- [Web development](https://www.techcelerant.com/services/web-development): Business websites, from template-based starter sites to custom builds.
- [Ecommerce development](https://www.techcelerant.com/services/ecommerce-development): Shopify and WooCommerce stores and custom commerce integrations.
- [SEO services](https://www.techcelerant.com/services/seo-services): Local SEO and monthly SEO plans.
- [PPC management](https://www.techcelerant.com/services/ppc): Google Ads, Microsoft Ads and Meta ads management.

## Key information

- [Pricing](https://www.techcelerant.com/pricing): Fixed-price packages and timelines for every service.
- [Case studies](https://www.techcelerant.com/case-studies): Project summaries, including a telemedicine app and a content management system.
- [Contact](https://www.techcelerant.com/contact): Project inquiry form and office details.

Some sites, mostly documentation sites, also publish an llms-full.txt file. Where llms.txt is a table of contents, llms-full.txt puts the full text of the docs into one file so an agent can read everything in one request.

How llms.txt compares to other web standards

Websites already use a few technical files at the root. Here is how llms.txt differs from them.

A robots.txt file governs access rights. If you disallow a directory in robots.txt, reputable crawlers will not visit it. An llms.txt file does not restrict access; it guides AI agents to the best available resources. To ensure your crawl rules are configured properly, review our guide on XML sitemaps and robots.txt.

Does your website actually need an llms.txt file?

Most don't need one yet. A few kinds of companies get real use from it now.

Businesses that should consider llms.txt now

• Software and SaaS companies. Developers ask AI coding tools how to use your software or API, and those tools often read documentation directly. An llms.txt file points them to current docs.
• Technical documentation sites and knowledge bases. Links to clean Markdown versions of each page make the docs easier for agents to read.
• B2B services firms that want to experiment. It is cheap to publish, though there is no evidence yet that it changes which vendors AI assistants recommend.

Businesses that can treat it as optional

Local retailers, restaurants and neighborhood service businesses don't need an llms.txt file. When AI assistants recommend local businesses, they draw on sources like Google Business Profiles, review sites and directories. Your time is better spent there, though publishing the file does no harm.

How to create and publish your llms.txt file

If you decide to publish one, it takes four steps.

Step 1: Select your essential pages

Leave out minor blog posts and tag pages. Choose the ten to twenty pages that matter most:

• Your core service or product pages.
• Your pricing or package overview.
• Your about page.
• Your contact or inquiry page.

Step 2: Write direct, factual descriptions

For every URL in your list, write a one-sentence description of what the page covers, without marketing language: what the service includes, typical prices, or who it is for.

Step 3: Provide clean Markdown versions if possible

Linking to normal HTML pages is fine. The proposal also suggests offering a Markdown version of each page at the same URL with .md added. Some documentation platforms do this automatically; for a typical business site it is a nice-to-have.

Step 4: Upload to your server root and test

Save your file as llms.txt (all lowercase) and upload it to your web server's public root directory (the same folder where your robots.txt and favicon live).

Open a web browser and visit https://yourdomain.com/llms.txt. Ensure the file displays as plain text without redirecting to an HTML page, returning a 404 error, or requiring authentication.

How llms.txt supports answer engine optimization

For most businesses, llms.txt is a small extra. What gets you cited in AI answers is the same work as good SEO: clear pages that answer questions, accurate structured data, and reviews and mentions on other sites.

For a broader look at how conversational search engines extract content, read our foundational guide on what is answer engine optimization and our specific playbook on how to get your business recommended by ChatGPT.

If you want a team to handle that work, our Growth SEO package at $1,499 per month includes a technical audit and fixes, a content plan, 4 optimized articles or pages a month, link building and AI search visibility tracking. See our SEO services for the details.

Immediate next steps for your website

If you run a software product with public docs, draft a twenty-line llms.txt file: your product name, a one-paragraph summary, and links to your getting-started guide, API reference and pricing page. Upload it to your site root and open the URL to check it loads as plain text. If you run a local service business, put that time into your Google Business Profile instead.