Does a practice website need an llms.txt file?

Every few months a new file shows up on the checklists that get sent to practice owners, and this year it’s llms.txt. It turns up as a must-have, usually with no word on which assistant reads it. So I went and looked at who does.

What is an llms.txt file?

An llms.txt file is a plain Markdown file placed at the root of a website, at /llms.txt, that gives language models a short summary of the site and a list of links to its most useful pages. Jeremy Howard proposed it in September 2024. It is a community proposal, and no search engine or assistant is obliged to read it.

The format is simple: a title, a one-paragraph summary, then sections of links with a line of description each. The proposal says it was meant mainly for inference, meaning a model reading your site while answering a question, rather than for training. Its best-known users are software companies. OpenAI, Anthropic and Google publish llms.txt files for their developer documentation, which is a sensible place for one: long technical docs that a coding assistant needs to read quickly.

A dental or medical practice website is a different animal. It has perhaps twenty pages, most of them short, and the facts that matter (who you are, where you are, what you treat, which insurance you take) are already on those pages in plain HTML.

Does Google use llms.txt for AI Overviews or AI Mode?

No. Google’s guide to its generative AI features says site owners do not need machine-readable files, AI text files or Markdown to appear, because Google Search does not use them. It adds that keeping an llms.txt file is fine for other services, and that it will neither harm nor help your visibility or rankings in Google Search.

Google added that note on 15 June 2026, and its changelog gives the reason: to answer questions from the community. When a search engine updates its documentation specifically to tell people a file does nothing for its results, the questions were coming in at volume. The same guide files llms.txt with chunking content and chasing inauthentic mentions, as tactics site owners can ignore for Google Search. Read that list before signing any proposal that leads with them.

Do ChatGPT, Claude or Perplexity read llms.txt?

Almost never, going by the only large log study published. Ahrefs checked 137,210 domains in May 2026 and found 97% of the llms.txt files they published got no requests at all. Of the requests the rest received, AI assistant bots such as ChatGPT-User made up 2.5% and AI search bots such as OAI-SearchBot 1.1%. That is a vendor study.

Two details from that study matter more than the headline. First, no AI bot went looking for an llms.txt file on sites that didn’t have one, so the bots aren’t checking for it by default. Second, the sample is Ahrefs Web Analytics customers, a group far more SEO-minded than the average practice, and 28% of them had published the file. Adoption among people who follow this closely is high. Use by the systems that write answers is close to nil.

Can an llms.txt file hurt a practice website?

Not directly. Google says the file neither harms nor helps in Search. The real risk is drift: a second summary of your practice that nobody remembers to update. If it lists an old address, a retired service or a provider who left, you have created one more place where your details disagree, and disagreement is what makes assistants hedge.

There is one small technical note. Chrome’s Lighthouse tool now includes an llms.txt check among its agentic browsing audits, and it flags a page when the server returns an error for the file. If you publish one, make sure it loads. If you don’t, that audit isn’t a reason to add one.

What should a practice do instead?

Spend the hour on the things assistants demonstrably use. Make sure the retrieval crawlers can reach your pages, keep your Google Business Profile complete and current, and make your name, address, phone, hours and services agree everywhere they appear. Those decide whether a model can find you and repeat you with confidence.

1. Check that OAI-SearchBot, PerplexityBot, Claude-SearchBot, Googlebot and Bingbot are allowed in robots.txt and are not blocked by your host’s bot protection.
2. Open your Google Business Profile and fill every field that applies to a practice, including services and hours for holidays.
3. Compare your details on your site, your profile, the insurance directories and the two or three medical directories that rank for your name. Fix every mismatch.
4. Only then, if your website platform generates an llms.txt file automatically, let it, and add a reminder to review it when anything above changes.

This file may matter one day. A standard can go from ignored to expected quickly once one large assistant commits to it. Nobody has, and the page is dated so you can see when this was last checked. I’ll revisit it when someone who writes answers says they read it.