Short answer: no.
If you run a normal business website, an llms.txt file won’t get your company mentioned more often by ChatGPT, Claude, Perplexity, or Google’s AI answers. One large study found the file is almost never fetched,1 another found no citation lift,2 and Google says it ignores it.3
It won’t hurt you either, as long as it’s accurate. If your CMS or SEO plugin already generates one, leave it in place.
The exception is companies whose product is documentation or an API.
What the file is
It’s a plain text file at the root of a website, at yourdomain.com/llms.txt. Inside is a short description of what the site is, plus a list of links to its most important pages, written in Markdown.
Jeremy Howard proposed the idea in 2024.4 An AI system trying to understand your business could read one summary instead of picking through menus, cookie banners, and layout code to find the parts that matter.
It doesn’t block crawlers or control what AI companies do with your content. If you want to block them, that still happens in robots.txt.
Ahrefs’ server logs
Ahrefs pulled the server logs of 137,210 websites for May 2026 and found that 97 percent of published llms.txt files received zero requests that month.1
Among the 3 percent that got any traffic, much of it wasn’t AI. SEO audit tools were the biggest single category, and Slack’s link preview bot fetched these files more often than Perplexity’s crawler did.
The study also found that AI bots never went looking for the file on sites that didn’t have one.
Google’s position
Google’s guide to generative AI features in Search, updated in 2026, lists llms.txt in a mythbusting section of things site owners don’t need.3
The guide says publishing one will “neither harm nor help” your visibility or rankings, because Google Search ignores the file. Chrome’s Lighthouse tool agrees and marks a missing llms.txt as not applicable.5
Asked on Bluesky why Google publishes llms.txt and Markdown pages for its own developer docs, John Mueller said it isn’t done for search.6 Coding tools can read HTML fine, so the simplified versions are, in his words, “more of a temporary crutch, perhaps to save some tokens.” For non-developer sites he doesn’t think it makes much sense.
SE Ranking’s citation study
SE Ranking analyzed roughly 300,000 domains in November 2025 to test whether the file earns more citations in AI answers.2 Sites with the file weren’t cited more often than sites without it.
When the researchers removed llms.txt from their prediction model, the model got more accurate, so the file wasn’t predicting anything.
An assistant answering a question is working from a search index built by crawling real pages. A self-description carries less weight than the pages it points to.
Developer documentation
The clearest finding in the Ahrefs logs was that a coding agent, not a search bot, was the top reader of these files. Anthropic’s Claude Code out-fetched every AI search bot, assistant, and training crawler in the dataset.
Its engineering team points to llms.txt as the usual home for LLM-friendly documentation, and Anthropic publishes one for its own developer docs.7
The file helps if your customers open a coding assistant and ask it to work with your API, since it gives that assistant a curated list of your documentation.
Risks
If the file lists pages you deleted or services you stopped offering, anything that reads it gets a wrong picture of your business.
Ahrefs also found that the largest research crawler in its dataset identifies itself as a prompt injection survey. Agents read these files and act on them, so anyone with write access to the file has an obvious place to put instructions.
To limit that risk, keep the contents to plain links and short descriptions, and review changes to it like code.
Recommendation
Ordinary business site with marketing pages and a blog: skip it.
Your platform generates one automatically: leave it in place and check it for dead links now and then.
Documentation, an API, or a product developers integrate with: publish one and curate it by hand. Link to it from your docs so agents know it’s there.
Ahrefs, “We Analyzed 137K Sites: 97% of llms.txt Files Never Get Read,” June 15, 2026. ↩︎ ↩︎
SE Ranking, “LLMs.txt: Why Brands Rely On It and Why It Doesn’t Work,” November 7, 2025. ↩︎ ↩︎
Google Search Central, “Google’s Guide to Optimizing for Generative AI Features on Google Search,” last updated July 10, 2026. ↩︎ ↩︎
Jeremy Howard, “The /llms.txt file,” September 3, 2024. ↩︎
Chrome for Developers, “llms.txt,” Lighthouse documentation. ↩︎
John Mueller, Bluesky thread, May 20, 2026. ↩︎
Anthropic, “Writing effective tools for AI agents,” Anthropic Engineering. ↩︎
