# What llms.txt can and can’t do for your site

_2026-11-09_

`llms.txt` is a plain-text index that points an AI reader toward the pages that matter. It can make a site's structure easier to discover when a person or tool asks an agent to read it. It does not make the content public by itself, and publishing the file does not guarantee that a model will fetch or cite your site.

## Keep the file useful

Start with the product or project name, one factual sentence, and a few links to stable pages. Group related pages under short headings. Link to canonical URLs that return useful content without a login.

```md
# Example product

> A one-sentence description of the product.

## Docs

- [Getting started](https://example.com/docs/start)
- [API reference](https://example.com/docs/api)
- [Security](https://example.com/security)
```

Do not copy the whole site into this file. A short index is easier to keep accurate, and a focused index avoids making the reader guess which page is current.

## What it does not replace

`llms.txt` is not the same as:

- **robots.txt**, which expresses crawler access preferences;
- **a sitemap**, which lists URLs for search crawlers;
- **authentication**, which protects private content; or
- **server-rendered page content**, which makes a page readable without running scripts.

Treat the format as a useful convention, not a search ranking factor. Keep relying on normal links, accurate page metadata, and a clear access policy.

## Check the file after publishing

Confirm the file is available and its links work:

```sh
curl -fsS https://example.com/llms.txt
npx backwatch example.com --pillar readability
```

The Backwatch [llms.txt checker](/tools/llms-txt-checker) checks whether the file exists, follows the expected structure, and links to reachable pages. It cannot tell whether every AI product will use the file.

For platform-specific setup, see the guides for [Next.js](/guides/nextjs), [Cloudflare](/guides/cloudflare), and [nginx](/guides/nginx).