What Is llms.txt, and Do You Need One?

An llms.txt file is a plain-text page at the root of a website, /llms.txt, that summarises the site and links to its most useful pages, written for AI agents. It is a community proposal, not an official standard, and AI engines differ in whether they read it. You do not need one to be found, and adding one costs little.

What is llms.txt?

The idea was proposed by Jeremy Howard and published on llmstxt.org in September 2024. A website is built for people: menus, scripts, pop-ups, long layouts. An AI agent that wants a quick, reliable overview of a company has to work through all of that. The proposal is to offer a short, curated, plain-text entry point instead: who we are, what we do and where the important pages are.

The file is written in Markdown, a simple text format. The proposal describes this order:

1. A first-level heading with the name of the site or project. This is the only required part.
2. A short summary in a quote block.
3. Optional paragraphs with more detail.
4. Sections, each with a second-level heading, that list links in the form [page name](address), optionally followed by a short note.
5. An "Optional" section for secondary links that an agent can skip when it needs less.

What does an llms.txt file look like?

Here is a small example for a fictional company. The names and addresses are placeholders.

# Northfield Packaging

> Northfield Packaging makes custom cartons for food producers.

## Products

- [Custom cartons](https://example.com/cartons): The carton range and how to request a sample.
- [Request a quote](https://example.com/quote): The form to ask for a quotation.

## Company

- [About us](https://example.com/about): Who we are and where we work.

## Optional

- [News](https://example.com/news): Company announcements.

Our own file is at millisave.com/llms.txt, if you want to see a real one.

Do AI engines read llms.txt?

Support varies, and no engine has made it a rule. We checked the public statements of three vendors in October 2026:

• Google's guidance on AI features says you do not need new machine-readable files, AI text files or special markup to appear in them. A page must be indexed and eligible to be shown in Google Search with a snippet.
• OpenAI's crawler documentation explains how its crawlers react to robots.txt. It does not mention llms.txt.
• Perplexity's bot documentation also describes robots.txt controls and does not mention llms.txt.

That is not proof that nobody reads the file. Some agents and developer tools may fetch it, and engines change. It does mean you should not expect it to decide whether you are cited, and you should be suspicious of anyone who sells it as a ranking trick.

Do I need an llms.txt file?

No. It is optional. Whether it is worth adding depends on your site.

• It is a reasonable extra if you have many pages, documentation or products, and you want one clean list of the pages that matter. Writing the summary also forces you to say clearly what the company does.
• It is not worth much time if your site has five pages and good navigation. Your sitemap, your headings and your page text matter more.

The cost is a small file and a habit of keeping it current. The risk is an outdated file, so every time you add or remove an important page, update the list.

What matters more than llms.txt?

The basics do more for your visibility than any extra file:

• Pages that are indexed, with real text, one clear title and one main heading each.
• A sitemap and a robots file that lets the crawlers you want read your site.
• Pages that open with a direct answer, and headings written as the questions customers ask.
• Honest structured data, only for what is visible on the page.
• One consistent description of your company across your site and profiles.

We explain the full list in How to show up in ChatGPT and Perplexity answers, and our SEO and AI-engine visibility page describes what we set up on the websites we build.

How do you add an llms.txt file?

1. Write the file in a plain-text editor, following the order above. Use full addresses (https://...) for the links.
2. Describe each link in one short line. Say what the page is for, not how good it is.
3. Save it as llms.txt and place it at the root of your site, so that it opens at yourdomain.com/llms.txt.
4. Open the address in a browser to check that it shows plain text and that every link works.
5. Put a reminder in your calendar to review it whenever your site changes.

Do not copy another company's file, do not stuff it with keywords, and do not include claims that your website does not support. Everything in it should be true and visible elsewhere on your site.

How is llms.txt different from robots.txt and a sitemap?

• robots.txt tells crawlers what they may and may not read. It is the established way to allow or block automated visitors.
• A sitemap is a machine-readable list of all the pages you want indexed. It is aimed at search engines.
• llms.txt is a short, human-readable summary with a curated list of links, proposed for AI agents. It does not allow or block anything.

The three do different jobs. One does not replace another.

What is the short answer?

llms.txt is a proposed, optional, plain-text summary of your site for AI agents. Add one if you want a tidy overview that you keep current, and do it after the basics. Do not expect it to guarantee a citation in any AI answer.

When we build a website, we add an llms.txt file and tell you plainly that no AI engine has promised to rely on it. If you would like a site built with the groundwork included, see our website development service or ask for a quotation.