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

llms.txt is a plain Markdown file at the root of your site that tells AI assistants what your site is and where the important pages are. It was proposed in 2024, it takes minutes to add, and it costs you nothing. It is also not a web standard, no AI company guarantees it reads the file, and no measured ranking benefit exists.

Both of those things are true at once, and most articles about llms.txt only tell you the first half. Here is the honest version, so you can decide whether it is worth your time.

What does an llms.txt actually look like?

It lives at yoursite.com/llms.txt and it is Markdown, not code:

# Acme Pest Control

> Family-run pest control in Belgrade, servicing homes and restaurants since 2004.

## Pages
- [Cockroach treatment](https://acme.rs/cockroaches): what the treatment involves and how to prepare.
- [Pricing](https://acme.rs/pricing): call-out fees and per-treatment prices.
- [Contact](https://acme.rs/contact): phone, address, opening hours.

That is the whole idea. A short, machine-friendly summary of a site, in a place a crawler can find without guessing.

How is it different from robots.txt?

People mix these up constantly, and the difference matters.

robots.txt tells crawlers what they may access. It is a genuine convention, respected for three decades, and a crawler that ignores it is misbehaving.

llms.txt tries to tell assistants what your site means. It is a suggestion. A crawler that never fetches it is not doing anything wrong.

So: robots.txt is something you must get right, because getting it wrong makes you invisible. llms.txt is something you may add, because it might help and cannot hurt.

Does it actually help?

Nobody has shown convincing evidence that it does. There is no public confirmation from OpenAI, Anthropic or Google that llms.txt influences whether you are cited.

The reasonable case for adding it is cheapness, not proof: it takes a few minutes, the file is tiny, there is no downside, and if adoption grows you are already there. The unreasonable case is anyone telling you it will get you into ChatGPT answers. It will not, on its own.

Be sceptical of tools that charge a subscription to "optimise your llms.txt". It is a text file. You can write it by hand in a text editor.

What should you put in it?

Keep it descriptive rather than promotional. The file exists to explain your site to a machine, and marketing language actively gets in the way of that.

• Do list the pages a stranger would need to understand your business: services, pricing, contact, your best guides.
• Do write one honest sentence per link about what is on that page.
• Don't list every page. A sitemap already does that, and a wall of links tells an assistant nothing.
• Don't stuff keywords. "Best cheapest pest control Belgrade #1 rated" reads as noise to a model and helps nobody.
• Don't put anything in it you would not publish openly. It is a public file.

Does it replace a sitemap?

No, and they are not competing. An XML sitemap is a complete machine-readable list of your URLs for search engines. An llms.txt is a short human-readable summary for language models. Keep the sitemap; add llms.txt if you want to.

FAQ

Where exactly does the file go?
The root of your domain: https://yoursite.com/llms.txt. Not in a subfolder. On WordPress you can upload it to the site root over FTP or through your host's file manager.

Do I need llms-full.txt too?
Some proposals mention a longer variant containing full page text. For a normal business site it is unnecessary and quickly goes stale. Start with the short one.

Will it get me penalised by Google?
No. Google ignores files it does not use. There is no risk in publishing it.

How often should I update it?
When your services or key pages change — realistically once or twice a year. A stale file describing services you no longer offer is worse than none.

Make one in a minute

Our free page check tells you whether you already publish an llms.txt. Inside SerpCue, the AI Visibility screen builds one from the pages we actually crawled on your site — real titles and real descriptions, not invented ones — and you download it and upload it to your root. If we have not crawled your site yet, we say so rather than handing you an empty file.