★ FREE TOOL · RUNS IN YOUR BROWSER

llms.txt Generator.
Useful, cheap,
and not a ranking mechanism.

Build the Markdown summary of your site that some AI tools look for at /llms.txt. It takes twenty minutes and does no harm. It is also not how you get cited, and this page explains what actually is. Nothing you type is sent anywhere.

Your details

One link per line: label, then URL, then an optional note, separated by a pipe character. The note is the part that earns its place.
llms.txt
# Your organisation

---

Last updated: 2026-10-19

What llms.txt is, stated plainly

It is a proposed convention: a Markdown file at the root of your domain that summarises what your site is and points to the pages that best explain it, written for large language models rather than for people. The placement echoes robots.txt, which is where most of the confusion starts, because the two files do entirely different jobs.

robots.txt expresses permission. llms.txt expresses nothing at all about permission. It grants no access, blocks no crawler, and carries no obligation on anyone reading it. It is a summary document, and the only thing it can do is be useful to something that chooses to read it.

Whether anything does read it is genuinely unsettled. Some tools and crawlers look for it. No major engine has committed to honouring it, and none has published that it influences anything. Anyone telling you an llms.txt file is how you get cited by an assistant is overselling a convention that no engine has promised to use.

So why produce one

Two reasons, and the second is the better one. The first is cost: twenty minutes of work, no maintenance burden if your business is stable, no downside. On that basis alone it is a reasonable bet on a convention that may become more widely read.

The second is the exercise itself. Writing a single sentence that says what your organisation does, a short paragraph of genuine context, and a curated list of the ten or so pages that actually explain you is clarifying in a way that is hard to fake. Most teams that do it discover their homepage does not say any of it plainly, which is a more valuable finding than the file.

Curation is the whole discipline. A file listing every URL on the site is a sitemap in worse formatting. The value is in choosing twelve pages over four hundred and saying, in a sentence each, why those twelve are the ones that matter.

What actually determines whether you get cited

Since this page exists to be honest rather than to sell a file, the short version is worth stating. Google says there are no additional requirements to appear in its AI features and no special optimisations necessary. The things that decide whether an assistant can use your content are considerably more boring than a new file format.

Can it be fetched. Most AI crawlers do not execute JavaScript, so a client-rendered site can be invisible to them while Googlebot copes fine. Server-rendered content is closer to a prerequisite here than it is for ranking.

Can it be resolved. A model works out who you are from facts that agree with each other across your site, your profiles and third-party sources. Inconsistency is the problem, not absence of a summary file.

Can it be quoted. Research presented at ACM SIGKDD in 2024 tested nine content strategies and found that adding quotations, citing sources and including relevant statistics were the strongest levers, while keyword stuffing performed worse than making no changes at all. That is a content discipline, not a configuration step.

The fuller argument, with the numbers, is in SEO vs GEO. If you want the version about what changes in a working week, it is in AI SEO vs traditional SEO.

Writing one worth reading

Lead with a quotable sentence. The summary line should be the sentence you would want a model to repeat when someone asks what your company does. Write it as a statement of fact, not a tagline. Taglines do not survive paraphrase.

Give each link a reason. A bare list of URLs tells a reader nothing about priority. One clause per link, explaining what is on the page and why it matters, is what turns a list into a map.

Say what you are not. If your brand name collides with another company, or you are frequently confused with an adjacent category, say so explicitly. Disambiguation is one of the few things a file like this is genuinely well suited to, and it is the part most people leave out.

Date it, and keep it true. A stale summary confidently describes a company that no longer exists, and nothing reading it can tell. If you cannot commit to revisiting it when things change, the honest choice is not to publish one.

Ours is at /llms.txt if you want to see one in the wild, alongside a longer machine-readable facts page at /llm-info.

FAQ

llms.txt, asked and answered.

Want to know if AI search can actually read you?

Send the domain. The founder checks whether your pages are reachable and quotable by AI crawlers at all, which matters considerably more than this file, and replies within 48 hours.

Apply for review ↗

A proposed convention for a Markdown file at the root of a site that summarises what the site is and points to its most useful pages, intended to be read by large language models. It is modelled loosely on robots.txt in placement, but it grants nothing and blocks nothing. It is a summary, not a permission system.

Some tools and crawlers look for it; there is no guarantee any particular assistant does, and no major engine has committed to honouring it. Treat it as low-cost and speculative rather than as a mechanism. If someone tells you an llms.txt file is how you get cited by ChatGPT, they are overselling a convention that no engine has promised to use.

No. Google states plainly that there are no additional requirements to appear in AI Overviews or AI Mode and no special optimisations necessary, which includes machine-readable files. What actually decides whether you can be cited is whether your content is crawlable, server-rendered, credible and quotable.

At the root of your domain, served as plain text or Markdown at /llms.txt. On most hosts that means dropping it in your public or static directory. It needs no configuration, no header and no registration anywhere.

No, and this is the most important misunderstanding about it. It expresses nothing about permission. If you want to allow or disallow AI crawlers, that is done in robots.txt with the relevant user agents, and enforcement still depends on the crawler choosing to respect it. An llms.txt file is a summary for whoever reads it.

A one-line description of what you do, a short paragraph of context, and a curated set of links to the pages that genuinely explain you, each with a sentence saying why it matters. Curation is the point. A file listing every URL on the site is a sitemap with worse formatting and gives a reader nothing to prioritise.

Whenever the things it describes change, which for most sites is a few times a year. A stale llms.txt is worse than none, because it confidently describes a version of your business that no longer exists, and anything reading it has no way to know that.

For most sites, yes, on a cost basis: it takes twenty minutes and does no harm. The secondary benefit is more reliable than the primary one. Writing a tight description of what you do, who you serve and which pages matter is a genuinely clarifying exercise, and it usually reveals that your homepage does not say any of it clearly.