llms.txt is a plain text file that sits at the root of your website and gives AI systems a short, curated guide to what your business is and which pages matter most. For most local businesses the answer to "should we have one?" is yes — it takes under an hour, carries no real downside, and is cheap insurance if the engines reading it keep gaining ground.
Why does llms.txt exist?
The proposal came out of a simple observation: web pages are built for browsers, not for language models. A typical homepage is a pile of navigation menus, scripts, cookie banners, and layout markup wrapped around a modest amount of actual information. Humans filter that out without thinking. AI systems have to burn limited context on it, and they don't always find the parts that matter.
llms.txt — proposed by technologist Jeremy Howard in 2024 — flips the arrangement. Instead of making a model dig through your site, you hand it a clean, structured summary in Markdown: here's who we are, here's what we do, and here are the pages worth reading if you want details. Think of it as the inverse of robots.txt. Where robots.txt tells crawlers what to stay out of, llms.txt tells them what deserves their attention.
What goes in an llms.txt file?
The format is deliberately simple. It's a Markdown document with a handful of conventional parts:
- A top-level heading with your business name — the one required element.
- A short blockquote summary — one or two sentences saying what you do, where, and for whom.
- Sections of links, each with a one-line description: your services pages, service-area pages, pricing, FAQs, about page.
- An optional section for secondary pages a model can safely skip when context is short.
For a plumbing company, the whole file might run fifteen lines: the name, a summary ("Emergency and scheduled plumbing for homeowners in the northern suburbs"), and links to the services, service-area, and contact pages with a sentence of context each. Accuracy matters more than length. This is one of the few places on the web where you write directly for the machine reader, so skip the marketing adjectives and state facts: services, locations, hours, what makes you the right call.
Do AI engines actually read it?
Here's the honest part: adoption is uneven, and no major engine has publicly committed to using llms.txt in a way that changes its answers. Some AI crawlers request the file; whether it moves the needle is unproven, and anyone who tells you otherwise is guessing.
So why bother? Because the cost-benefit math is lopsided. Publishing one costs you an hour once and a few minutes a year to maintain. If it does nothing, you've lost almost nothing. If assistants increasingly consult it — plausible, given the direction of travel — you're ahead of most local competitors, who have never heard of it. That's the definition of cheap insurance. Just keep it in perspective: llms.txt is a courtesy map, not a strategy. The heavier lifting lives in making your content answer-shaped and making sure crawlers can reach your site at all.
How do you publish one on any site?
llms.txt is just a text file, so publishing is mostly a question of getting it to your domain root:
- Write the file in any text editor and save it as llms.txt.
- Upload it to the root of your site, next to robots.txt — via SFTP, your host's file manager, or your deploy process on a static site.
- WordPress users can upload through the hosting file manager, or use one of several plugins that generate and serve the file.
- Verify by loading yourdomain.com/llms.txt in a browser. If you can read it there, so can a crawler.
One caveat: some closed website builders don't let you place arbitrary files at the domain root. If yours doesn't, don't contort yourself over it — this is a nice-to-have, not a foundation.
What llms.txt won't do
It won't unblock a firewall that's turning AI crawlers away, it won't rescue thin content, and it won't polish your Google Business Profile — which still does a great deal of work in how assistants describe local businesses. Treat it as one small item on the larger generative engine optimization checklist: publish it, verify it, and move on to the work that demonstrably matters. If you'd rather see the whole checklist scored for your site in one pass, an automated audit like Speak Local's will check llms.txt alongside the things that actually gate visibility — crawler access, rendering, and structured data. Either way, spend the hour. Few line items in marketing cost this little and age this well.
