We publish a post called What Is llms.txt (and Do You Actually Need One)? that says, in plain language, that most businesses probably do not need one yet.
We still stand by that. It would be strange to sell you a service on a page that contradicts our own article, so this page starts where that one ended: with who this is genuinely for, and who should close the tab and keep their money.
The honest state of llms.txt
llms.txt is a proposed convention. It is a plain text file at the root of your site that hands AI systems a clean, structured summary of what you do and which pages matter, instead of making them infer it from your navigation.
As of this writing, no major AI crawler has publicly and formally committed to fetching or prioritizing it the way every search engine reliably parses robots.txt. Google has not adopted it. Anyone telling you ChatGPT definitely reads your llms.txt today and weights it when deciding who to cite is ahead of the evidence.
That could change, and standards sometimes get adopted quickly once one major player commits. It has not happened yet. Any page selling you this should say so before it asks for money.
So why would you pay for this
Because the file is not really the deliverable. The useful part is what writing it forces you to do.
A correct llms.txt has to state, in a few lines a machine can parse, what your business does, who it serves, and which twenty or thirty pages actually matter. Most sites cannot survive that exercise. Halfway through it you find three pages describing the same service in different words, a services page that never says which city you operate in, and eleven blog posts that would embarrass you if an AI assistant quoted them.
That inventory is the value. The file is the artifact it produces. If the file turns out to matter later, you already have it. If it never does, you still fixed the things you found on the way.
Who this is actually for
Do this if your site is large or sprawling, roughly dozens of pages upward, where a machine crawling the whole thing genuinely benefits from a map. Do it if you have already handled the structured data and crawler access work, because those are confirmed to matter today and this is not a substitute for either.
Do it if you publish a lot and want control over which version of your story an assistant picks up. And do it if you would rather have the file in place cheaply now than scramble if adoption arrives suddenly.
Who should not buy this
If you run a fifteen page site with clean navigation, real schema markup and content that plainly states what you do, you get very little from an llms.txt file that a five minute read of your site would not already tell a model. Take the copy and paste starter from our post, spend twenty minutes, and keep the rest.
If your structured data is missing or wrong, fix that first. It is confirmed to affect what AI systems cite, and llms.txt is not. Doing this before that is decorating a house with no foundation.
If someone has told you this is urgent, they are selling panic. It is one small, low cost signal, not a replacement for the structural work.
What we do
We write and install both files, and we check the things around them that actually determine whether an assistant can read you at all.
- A content inventory that decides which pages belong in the file and which do not
- A written llms.txt with a real summary of your business, not a list of URLs
- An llms-full.txt with the expanded content where it earns its place
- Crawler access verified for GPTBot, ClaudeBot, Google-Extended and PerplexityBot
- A robots.txt review, because blocking these by accident is more common than people expect
- A findings note on anything the inventory turned up that is worth fixing
Delivered in 5 business days. It is a small job and we price it like one.
What this will not do
It will not make ChatGPT recommend you. No file does that. Being recommended comes from being a source worth citing, which is a longer and harder project.
It will not affect your Google rankings. llms.txt is not a search engine signal and we have seen nothing credible suggesting it will become one.
It will not produce a measurable result you can point at next month, because there is currently no reporting surface for it. If you need proof that something moved, an AI visibility audit measures whether assistants name you, and that is the number worth tracking.
Where this fits
This is one line item inside AI search readiness and AEO, which is the broader work: structured data, entity signals, answer shaped content, and the technical access that decides whether any of it is visible. If you are trying to get cited by AI assistants, that is the project. This is a small, cheap, sensible piece of it.
You can also start with our free website audit tool, which checks crawler access and structured data in about a minute and will tell you whether the foundation is in place before you spend anything.
Request a quote and we will tell you what it costs, and whether we think you need it. If the answer is that you do not, we will say so.