An llms.txt is a plain-text file placed at the root of a website that gives AI assistants a concise, structured summary of what the site contains and which pages matter most. It is an emerging convention — not a ranking factor and not a requirement — but it is free to add and takes about ten minutes.

What llms.txt actually is

The file lives at https://yoursite.com/llms.txt. It is written in a simple Markdown-like format intended to be read by language models, not by people. Think of it as a table of contents for an assistant that wants to understand your site quickly instead of parsing dozens of HTML pages.

The convention was proposed as a way to give language models a clean starting point: a short description, a curated list of the most important pages, and optional details. The idea is that a model can read one small file and immediately know where to look.

There is no public evidence that any major answer engine requires llms.txt to cite a site, and no evidence that it is a direct ranking signal. Its value is convenience and clarity. Add it because it is cheap and tidy, not because it will single-handedly change your visibility.

A complete copy-paste example

Here is a working file you can adapt:

# Your Company

> A one- or two-sentence description of what you do, who you serve,
> and what makes your pages worth citing.

## Key pages

- [Homepage](https://yoursite.com/): overview of the company and what we offer.
- [Pricing](https://yoursite.com/pricing/): current plans and costs.
- [Documentation](https://yoursite.com/docs/): technical reference and guides.
- [Blog](https://yoursite.com/blog/): articles on industry topics.

## Optional

- [Full documentation](https://yoursite.com/llms-full.txt): complete text of our docs.

Keep it short. A few dozen lines covering your genuinely important pages is enough. Longer files defeat the purpose of a quick summary. Some implementations also publish an llms-full.txt with expanded content, but that is an optional extension, not part of the basic convention.

Format basics

  • Start with an H1-style line (#) naming your site or organisation.
  • Follow with a blockquote (>) containing a short summary.
  • Add one or more sections (##) grouping links, commonly Key pages and Optional.
  • List each page as a Markdown link with a short description after a colon.
  • Use absolute URLs so a model never has to guess the domain.

There is no formal specification enforced by any standards body, so treat this as a widely used pattern rather than a strict protocol.

Where to put it and how to add it

The file must be served from the root so it resolves at https://yoursite.com/llms.txt and returns HTTP 200 with a plain-text content type.

  1. Create a file named llms.txt.
  2. Paste the structure above and replace the content with your own.
  3. Upload it to the root of your site.
  4. Open the URL in a browser and confirm it loads.

Platform specifics:

  • Cloudflare Workers / Pages or other static hosts: add the file to your assets or public directory and redeploy.
  • WordPress: place it in the site root via FTP, your host's file manager, or a plugin that manages root files. Some SEO plugins now include an llms.txt option.
  • Shopify: arbitrary root files are restricted; you typically need a workaround or a proxy route.
  • Webflow / hosted builders: check whether custom root files are supported; many require a custom-domain workaround.

If your site is generated by a build step, generate llms.txt from the same source of truth as your sitemap so the two never drift apart.

How to validate it

  • Open https://yoursite.com/llms.txt and check the text renders as plain text.
  • Confirm every listed URL resolves and is not blocked by robots.txt.
  • Check the response is HTTP 200, not a redirect chain or a 404 page styled as text.
  • Use the free checker at https://ai.ynnadmusic.uk, which reports whether an llms.txt is detected on your site.

Limitations you should know

  • It does not override robots.txt. If a crawler is blocked there, an llms.txt will not unblock it. Fix access first.
  • Support is inconsistent. Which assistants, if any, actually read these files is not publicly documented in a stable way and may change.
  • It is not a substitute for good pages. A clean summary of thin, unstructured pages is still a thin site. Clear, crawlable, answer-first content matters more.
  • It is not a guarantee of citation. Nothing you put in this file forces an assistant to quote you.

Treat llms.txt as one small, low-cost signal among many. The higher-impact work is crawler access and content clarity; see the 12-point AI visibility checklist for the full picture.

Frequently asked questions

Is llms.txt a ranking factor?

There is no public evidence that it is. It is best understood as a convenience file that may help a model find and summarise your most important pages. It costs almost nothing to add, so the upside is worth taking even though the effect is unproven.

Will llms.txt stop AI crawlers from being blocked?

No. Crawler access is controlled by robots.txt and related directives. Confirm crawlers can reach your site before adding an llms.txt. See the GPTBot and robots.txt guide.

How long should llms.txt be?

Short. A few dozen lines covering your best pages, a short description, and optional links is enough. If it becomes a full copy of your site, it has stopped being a summary.

Do I still need a sitemap if I have llms.txt?

Yes. They serve different purposes. A sitemap is a machine-readable list of URLs for search engines; llms.txt is a curated human-and-model-readable summary. Keep both, and generate them from the same page list where possible.