Skip to content
codecollab.pl
§ Blog llms.txt

llms.txt - what it is and whether it is worth having

A small text file at the root of your site tells AI who you are and what on your site is actually worth reading

Krystian Kacik 7 min read
Contents

A small text file at the root of your site tells AI who you are and what on your site is actually worth reading. It sounds unremarkable, but that is exactly how llms.txt works - you upload it to yourdomain.com/llms.txt and from then on ChatGPT, Perplexity or other AI tools have a ready-made guide to your company to reach for, instead of guessing from scattered subpages. If they want to, because none of them has committed to it - and I will come back to that. I rolled this file out on this very site (you can look at codecollab.pl/llms.txt), so I write as a practitioner, not a theorist.

What llms.txt is, in plain language

It is an ordinary text file sitting in the main directory of your site, next to the other technical files. There is no code inside - just a short, orderly description: who you are, what you do, which subpages matter most and what a reader will find on them. Formally it is a Markdown file, meaning text with simple formatting: the company name as a heading, one summary sentence, then sections with lists of links and a short note next to each. That is all.

What is it for? AI tools increasingly answer questions like “who can build me an online store” or “how much does fixing a website cost”. To answer, they first have to understand your site - and sites tend to be cluttered: menus, banners, pop-ups, content broken across dozens of subpages. llms.txt is the cheat sheet that says outright: this is what matters, here is how to read it.

llms.txt and robots.txt - similar name, opposite job

If you have ever heard of robots.txt, the analogy is simple. robots.txt decides whether a bot gets in at all - and where it is not allowed to look. llms.txt does not decide about access at all: it only points out what is worth reading and how to organise it in your head. One file is the doorman, the other is the map at the entrance. Both can (and should) exist on a site side by side, because they do entirely different things.

Where it came from and why it is a bet

llms.txt is a proposed standard from 2024 - not an official requirement, just an idea catching on from the ground up, because it is simple and harms nobody. Honestly: no AI tool has to read this file, and as of today no major one has committed to it. Google said outright that it does not support llms.txt, and OpenAI and Anthropic point to robots.txt in the documentation for their crawlers. Server logs show that the file is mostly fetched today by SEO audit tools and coding assistants, not by ChatGPT or Perplexity. The cost of rolling it out, on the other hand, is laughably small - it is one text file.

So I treat it as a low-stakes bet with a potentially large payoff. If the standard does not catch on, you lost a moment of work. If it does, your company is described in AI answers in your own words rather than whatever the algorithm scraped together from the site.

What to put inside

A good llms.txt is short and concrete. I stick to this skeleton (the live example is the file at codecollab.pl/llms.txt - you can open it and see how it looks in practice):

  1. Who you are - one or two sentences positioning the company: what you do, for whom, what sets you apart
  2. Service areas - with the terms and prices you want AI to know; if you quote individually, say so outright
  3. Key subpages - a list of links with a short note on what each contains
  4. Contact - how to reach you and how working together starts

There is one rule: you write for a reader who does not know your industry but understands orderly text perfectly. No marketing ornament - AI will filter it out anyway.

llms-full.txt - the extended version

Alongside the short guide there is a variant called llms-full.txt: the full content of your most important pages rewritten as clean text, without menus, footers or ornament. The short file tells AI where to look - the full one hands over the whole content on a plate. Both files are live on this site and I treat them as a set: llms.txt as the table of contents, llms-full.txt as the reading itself.

How to create llms.txt on WordPress

You have two routes:

  • By hand - you write the file in a plain text editor and upload it to the server over FTP or the hosting file manager, into the site’s main directory. Maximum control over the content, zero dependency on add-ons.
  • With a plugin - there are SEO plugins and standalone generators that assemble llms.txt automatically from your site content. Faster, but the result can feel mechanical: the plugin does not know what matters to you, so it treats everything equally.

Whichever route you take, one thing matters more than how the file was created: keeping it current. An llms.txt with an outdated offer is the shortest path to AI telling your customers about prices and services you no longer have.

What helps: add llms.txt to the checklist for every change on the site - a new service, a price change, new contact details. You change the offer, you change the file, the same day. An out-of-date guide is worse than none, because it feeds AI old information with your signature on it.

Does llms.txt actually do anything

The honest answer: there is no guarantee, because the standard is young and no AI tool promises to take it into account. But there is a second layer. The foundations of visibility in AI answers - which I cover more broadly in the piece on SEO and AI - require tidying up your content anyway: a clear description of services, a sensible subpage structure, concrete answers to customer questions. llms.txt is the natural summary of that work. If you have nothing to write it from, that is a signal the problem runs deeper than a missing file - and that is exactly what a proper SEO audit surfaces.

On my side llms.txt is live and I include it in every rollout where it makes sense - alongside the rest of the visibility work I describe in the section on SEO and AI visibility. And if you want to know where you stand: as part of the free diagnosis I also check the site’s AI readiness - whether there is anything to build such a guide from, and what AI says about you today.

Frequently asked questions

What is llms.txt? It is a text file in the main directory of your site (yourdomain.com/llms.txt) that describes your company to AI tools: who you are, what you do and which subpages matter most.

How does llms.txt differ from robots.txt? robots.txt decides about access: it lets bots in or closes parts of the site to them. llms.txt blocks nothing - it only points out what is worth reading and how to understand it. Two separate files with different jobs, and they can work side by side.

Is llms.txt mandatory? No. It is a proposed standard from 2024, voluntary and for now without official support from the large AI providers. A site without the file works normally and can still be cited by AI - llms.txt is a convenience, not a pass.

How do I check whether my site has llms.txt? Type your site’s address in the browser with /llms.txt on the end. If you see orderly text about the company, it is there. An error or a blank page means it is not.


Want your company to show up in Google and in ChatGPT answers? I work on technical SEO and AI visibility from the code side, not the reporting side. Tell me which questions matter to you and I will quote it.

§ Quote in 24h

Facing a similar problem and not sure where to start?

Describe the scope in two sentences or send a link. I tell you what to fix first, and you get a fixed bid in writing within 24 hours - no "from X" pricing.

Send your scope - quote in 24h