Does llms.txt actually do anything?

The short answer

llms.txt is a plain-text file you put at yourdomain.com/llms.txt. The idea: give AI tools a clean map of your site and a curated list of your most important content, written for language models rather than for a search crawler.

Does it do anything? It depends entirely on who you are hoping will read it.

For Google's AI features (AI Overviews and AI Mode): no. Google has said plainly that you should not create an llms.txt file for it. Those features pull from the regular search index, so the file is wasted effort there.

For the wider AI ecosystem (some ChatGPT integrations, Perplexity, various agents and scrapers): maybe. It is an emerging convention with patchy adoption. Some tools look for it, most do not yet, and nobody is contractually obligated to.

So the honest position is: llms.txt is a low-cost, low-risk bet on a convention that might pay off across part of the AI landscape, and definitely does nothing for the biggest player. That is worth understanding before you spend an afternoon on it.

What Google actually says

Google's AI features optimization guidance is direct on this point. Its mythbusting advice tells site owners not to create llms.txt files or custom AI-specific markup for Google's AI features.

The reasoning is consistent with how Google's AI works. AI Overviews and AI Mode use retrieval over the core Search index, plus a query fan-out technique that fires off related queries at once. They draw from the same index that ranks blue links. There is no separate AI pipeline reading a separate file. If your page is crawlable, indexable, and helpful, it is already eligible. A llms.txt file does not add a new door, because Google is not knocking on that door.

This is the same message running through Google's guidance generally: "optimizing for AI" is, for Google, just SEO. Be indexable, be helpful, be well-structured. That is the path in.

So if your entire goal is Google's AI answers, skip llms.txt and spend the time on the fundamentals in our guide to getting found by AI search instead.

Why the file still exists, and who it is for

Here is the nuance that gets flattened in most takes on this topic. Google is not the whole answer-engine ecosystem.

llms.txt was proposed as an open convention, not a Google standard. The intent is to help any AI tool quickly understand a site without crawling and parsing every page: a short Markdown file that says what the site is, who it is for, and links to the canonical versions of your key pages and docs. Think of it as a hand-drawn map you leave at the front door for whoever happens to want one.

Adoption is real but uneven. Some AI tools and agents look for it; many still ignore it entirely. It is not an official standard backed by a major search engine, and no AI provider promises to honour it. That is the awkward truth: it might help with some non-Google tools, and it might be quietly ignored by all of them. Both outcomes are common today.

That is why we treat it as an optional extra, not a requirement. If you want the definition and the current state of adoption, our glossary entry on llms.txt keeps it up to date.

When it is worth doing

You will get the most out of an llms.txt file if:

• You have a docs-heavy or content-heavy site. Developer tools, SaaS products, and knowledge bases benefit most, because the file can point AI tools straight at your canonical docs instead of letting them guess.
• You publish content you genuinely want AI tools to summarise correctly. A clean map reduces the odds an agent reconstructs your offering from a stale cached page or a third-party mention.
• You can keep it accurate. A map that points at dead URLs is worse than no map. If you cannot maintain it, do not ship it.

It is a thirty-minute job, not a project. The risk of having one is close to zero. The upside is modest and concentrated on non-Google tools. That is the trade you are actually making.

What llms.txt is not

It is worth being clear about the things llms.txt will not do, because the hype tends to oversell all of them.

• It is not a ranking signal. No search engine ranks you higher for having one.
• It does not get you into Google's AI answers. Covered above. The index does that.
• It does not replace good HTML. If your pricing is trapped in an image or your main content is rendered in a way crawlers cannot read, a text file does not rescue you. Real, selectable text on the page still wins.
• It is not a substitute for being helpful. Google's wider guidance on creating helpful content applies to AI surfaces too. A tidy map of thin content is still thin content.

Treat llms.txt as a small, optional courtesy to part of the ecosystem, not a lever that moves your visibility on its own.

The pragmatic move

If you want one, write it, keep it accurate, and move on. Do not let it crowd out the work that actually matters: clean HTML, readable structure, fast pages, and content made to help the person asking the question.

We do not ship an llms.txt generator, and we are not going to build a funnel around a file we cannot show works. It is a short Markdown file — you can write it by hand in a few minutes if you want one. Then go back to the fundamentals, which is where almost all of your AI visibility is won or lost anyway.