What is an llms.txt file, and do I need one?
Short answer
An llms.txt file is a proposed convention: a plain text file at the top level of a website that summarizes the site and links to its main pages for AI language models. You don't need one. There's no reliable evidence the major assistants act on it. It does no harm, and it doesn't replace clear pages.
An llms.txt file is a proposed plain text file that summarizes a website for AI language models, and no, a local business doesn’t need one. It’s a proposal, not an adopted standard. I know of no reliable evidence that the major assistants read it or act on it. Adding one does no harm if it’s accurate, and it isn’t a substitute for pages that state your facts clearly.
What the file is
The idea was put forward in 2024: a short summary of your site written for a machine reader. The file sits at the top level of the site, at yourdomain.com/llms.txt, and is written in Markdown, a simple way of formatting plain text. It holds the name of the site and a brief description, followed by a list of links to the most useful pages with a line about each.
The reasoning behind it’s fair. Web pages carry menus, scripts and layout code that get in the way of the actual content, so a clean summary could make a site easier for a language model to use.
Whether anything reads it
As far as I can tell, not in a way a business can count on. A convention like this only becomes useful once the systems it’s aimed at adopt it. To my knowledge, none of the major assistant makers has committed to using llms.txt when writing answers. This could change. If it does, the company concerned will say so in its own documentation, and that is the source to trust over a marketing claim.
You can also look for yourself. Your server logs record every request, including any request for that file, and your host or developer can tell you whether anything has asked for it.
How it differs from robots.txt and a sitemap
These files are often mentioned together and they do different jobs.
| File | What it is for | Status |
|---|---|---|
| robots.txt | Tells crawlers which parts of the site they may fetch | Long established and followed by well-behaved crawlers |
| XML sitemap | Lists your pages so search engines can find them | Established and supported by the main search engines |
| llms.txt | Summarizes the site for language models | Proposed, with no confirmed use by the major assistants |
An llms.txt file neither grants nor denies access. Putting a page in it doesn’t invite a crawler, and leaving a page out doesn’t keep one away. Access is controlled elsewhere, in the places you’d go to block AI crawlers from your website.
Whether to add one anyway
You can, as long as you expect nothing from it. The cost is small. The risk is a stale file. It becomes one more place where your services, hours and cities are written down, and it’s easy to forget when those change. A summary that says you serve Carlsbad and Encinitas after you have stopped taking jobs there’s worse than no summary. If you add one, keep it short and put it on the list of things to update.
Be wary of a proposal that lists an llms.txt file as a deliverable that will get you recommended. Nobody can show that it does.
What to do with the same effort
The pages themselves are what assistants are known to read, so work on those. That means business facts written as plain text in the HTML, one question per page with the answer in the first sentence, structured data that matches what the page says, and the same details on your listings elsewhere. The writing side is how a page should be written so an AI can quote it, and the service that covers this work is LLM SEO (LLM stands for large language model).