Draft your llms.txt from your sitemap.
Enter a domain. We find the sitemap your site declares, read it, and hand back a draft you can edit, with the two lines that matter left blank on purpose.
We do not store the sites people check.
Anyone who has been told they need one.
An llms.txt is a plain text file at the root of your site that says what your business is and which pages answer which question. It takes about ten minutes and costs nothing.
It is also being sold as essential by people who know better. Before you pay anyone for one, read the honest section below, then take the draft this tool gives you and edit it yourself.
Six things, and two of them are refusals.
It reads your sitemap
You do not type your pages in. We find the sitemap your site declares in its own robots.txt, read it, and list what is actually published, including the pages you have forgotten about.
It groups what it finds
Pages whose names read like questions go under one heading, free tools under another, legal pages last. The grouping is the part a reader uses.
It leaves the summary blank
On purpose. That one sentence is the line an engine reads first, and it is the line we refuse to guess, see below.
It leaves the contact block blank
Your address and phone are not in a sitemap. Guessing them would be worse than omitting them, and they must match your Google Business Profile exactly.
It tells you if you already have one
If your site already serves an llms.txt, the draft is a comparison. A file a person wrote is almost always better than a generated one.
It caps a large site
Past 120 pages we stop and say so. An llms.txt is a map for a reader, not a second sitemap.
No major assistant has confirmed it reads llms.txt.
It is a proposed convention from 2024, not a standard anyone has agreed to honour. We have read the crawler and indexing documentation that OpenAI, Anthropic and Google publish, and none of it commits to reading this file. If that changes we will say so here.
So anyone telling you an llms.txt is required for AI visibility is selling you something. The claim that sells is “AI reads this file”. The true claim is weaker and duller: it costs almost nothing, and if the convention is adopted you already have one.
We publish one ourselves — at ultimateelitesolutions.com/llms.txt — for exactly that reason, and for one more. Writing it forces you to say in a single sentence what your business does and where its answers live. That is worth doing whether or not a machine ever reads the result.
If you want the things that demonstrably do decide whether an assistant can see you, they are crawler access and discovery, and both are free to check.
The honest limits.
Whether anything reads it
The honest answer is that no major assistant has publicly confirmed it does. See the section below. It is the most important thing on this page.
Whether your summary is true
We leave it blank, so what goes there is yours. Make sure it agrees with your address: our own file contradicted itself on exactly that point.
Pages not in your sitemap
If a page is missing from the sitemap it is missing from the draft. Our sitemap validator will tell you what is absent.
Whether the pages are any good
A tidy index of thin pages is a tidy index of thin pages. The file helps a reader find your answers; it does not make them answers.
Five moments worth ten minutes.
When you first hear about llms.txt
Read the honest section below before you spend anything on it. It is cheap to add and it is not a strategy.
After publishing a batch of pages
Regenerate and compare. The diff shows what you have added since you last looked at your own site as a whole.
When your site map is a mess
The grouping alone often shows you the structure you meant to have and do not.
Before paying anyone to write one
It is a text file. See what a generated one looks like before you buy a bespoke one.
When you want to see a competitor's
Try their domain. If they publish one, you will see exactly how they describe themselves to a machine.
Most of them are a blank form.
A blank form asks you to type your pages in. Your pages are already listed in your sitemap. Reading it is the whole job, and it finds the pages you forgot you had.
Most will also write your summary for you. Ours will not, and that is not laziness. Our own summary line claimed a city this business had left while the address in the same file said otherwise, a contradiction that sat live until we happened to open the file. A generator willing to write that sentence would have shipped the same mistake to every site that used it.
And most imply the file is read by AI. We have said plainly above that no assistant has confirmed it. You can judge the rest of this site by whether that sentence is still here when it stops being convenient.
A map does not change the territory.
An llms.txt points a reader at your answers. It does nothing about whether those answers exist, whether they match the questions customers actually ask, or whether the facts about your business agree across every source an engine can reach. That is the work.
We publish how we test it, including the answers where we are not named, on our methodology page. Prices are on the pricing page.
What owners ask about this.
Does ChatGPT or Claude actually read llms.txt?
No major assistant has publicly confirmed that it does. It is a proposed convention from 2024, not a standard anyone has agreed to honour, and no vendor documentation we have read commits to reading it. Anyone telling you it is required for AI visibility is selling something. It costs nothing to publish and it may matter later, which is a different and much weaker claim than the one usually made.
Then why publish one at all?
Three honest reasons. It costs about ten minutes. If the convention is adopted you already have one. And writing it forces you to say in one sentence what your business does and which pages answer which question, which is worth doing whether or not a machine ever reads the result.
Why will the tool not write my summary?
Because we got that exact line wrong on our own site. Our llms.txt said the business was based in one city while the address further down the same file named another, and that contradiction sat live for days. A generator confident enough to write that sentence from a sitemap would ship the same error to everyone who used it. A blank you must fill beats a sentence a machine guessed.
Where does the file go?
At the root of your site, served as plain text: yourdomain.com/llms.txt. Add a line to robots.txt pointing at it, the same way you point at your sitemap.
Is llms.txt the same as robots.txt?
No, and they do opposite jobs. robots.txt says what a crawler may not fetch. It is enforced and it is read by everyone. llms.txt tries to say what your site is about and where the useful pages are, and nothing is obliged to read it.
How long should it be?
Short enough to read. A sentence about the business, a handful of sections, and the pages that answer real questions. If it approaches the length of your sitemap you have built a second sitemap, which nothing needs.
Will this fix my AI visibility?
No. Being named by an assistant depends on your pages answering the questions customers actually ask, and on the facts about your business agreeing across every source an engine can reach. An llms.txt is a map; it does not change the territory.
Do you store the sites people check?
No. The draft is built, returned to your browser, and nothing is kept. There is no sign-up and no email field on this page.
Related: Sitemap validator · Robots.txt checker · Structured data validator · Canonical tag checker
By Steven Allbright, founder, Ultimate Elite Solutions · Last updated: 2026-10-08
The next checks worth running.
Each one takes seconds and needs no sign-up. They fail in order, so start at the top.
Title tag checker · AI crawler block checker · Robots.txt checker · How we run a real test · Get the free report