SEO Strategy Tools

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

By Cécile, Impact SEO Consulting · July 12, 2026

If you have spent any time reading about AI search this year, you have probably seen llms.txt come up. The pitch is simple: it is a single file that tells AI assistants like ChatGPT, Claude, and Perplexity which pages on your site matter most. The reality is a little more nuanced, and worth getting right, because in May 2026 Google started checking for it inside Lighthouse’s agentic-browsing audit. That turned llms.txt from a nice-to-have into a signal that shows up in tooling founders and developers actually look at.

Here is what it is, what it is not, and how to publish one that pulls its weight.

What llms.txt actually is

llms.txt is a plain Markdown file you publish at the root of your domain, at https://yourdomain.com/llms.txt. It is a curated map of your site written for machines: a short title, a one-line description of what you do, and a set of linked sections pointing at the pages you most want an AI to read and cite.

Think of it as the opposite of your XML sitemap. A sitemap lists every URL you have so search crawlers can discover them all. An llms.txt lists only the handful of pages that best represent you, so an AI assistant with limited context can find your strongest material fast.

A minimal one looks like this:

# Acme Analytics

> Self-serve product analytics for small SaaS teams.

## Product
- [How Acme works](https://acme.com/product)
- [Pricing](https://acme.com/pricing)

## Guides
- [Set up your first funnel](https://acme.com/guides/funnels)

What it is not

Two honest caveats, because the space is full of hype.

First, llms.txt is not a ranking factor in classic Google search, and there is no guarantee any given AI system reads it today. Adoption is still uneven. What has changed is that publishing one is now cheap, low-risk, and increasingly expected by the tools that audit AI-readiness.

Second, llms.txt does not grant access. If your robots.txt blocks GPTBot or ClaudeBot, or your page carries a noindex, then no amount of llms.txt will get you cited. Access comes first; curation comes second. It is worth checking that your key pages are even reachable by AI crawlers before you worry about this file at all.

What belongs in it

The single most common mistake is dumping your entire sitemap into llms.txt. That defeats the purpose. The value is in the curation. A good llms.txt is short, and every link earns its place.

Include:

Leave out:

If a page would not make your own shortlist of “read these five things to understand us,” it does not belong in llms.txt.

How to publish one

  1. Draft the file. You can write it by hand, or generate a first pass from your sitemap and trim it down.
  2. Save it as llms.txt and upload it to the root of your site, so it resolves at https://yourdomain.com/llms.txt.
  3. Confirm it loads as plain text, not as an HTML page. Some single-page apps serve their app shell for unknown paths, which silently breaks the file.
  4. Revisit it when you publish something genuinely worth surfacing. This is a shortlist, not a changelog.

The honest bottom line

llms.txt is low-effort insurance for AI search visibility. It will not transform your traffic on its own, and anyone selling it that way is overpromising. But it is quick to publish, it signals that you take AI readiness seriously, and it is now baked into the tooling that judges your site. If your best pages are already reachable by AI crawlers, adding a clean, curated llms.txt is an easy next step.

If you want a head start, our free generator reads your sitemap, drafts a sectioned llms.txt for you, and lints any file you already publish. Curate the draft it gives you, and you are done in a few minutes.

Try the llms.txt Generator & Validator

Draft an llms.txt from your sitemap and lint the one you already have.

Open the free tool →

← All posts