Does llms.txt Help SEO or AI Visibility?

Key takeaways

The short version of this page, in the order the sections below develop it. Each point is stated here in full so it can be read without the rest of the page, and the section that argues it gives the measurement and the source behind it.

• Google states that Search, including its generative AI features, ignores llms.txt: the file neither harms nor helps.
• OpenAI, Anthropic and Perplexity document robots.txt as the control for their search crawlers; none documents llms.txt as an input to what gets cited.
• The two public measurements we read found no effect; both come from vendors that sell AI-visibility software and neither is independently reviewed.
• The file is worth publishing when AI agents read your developer documentation, and it should be generated, not hand-written.

Does llms.txt help SEO or AI visibility?

No for Google Search, and nobody has shown that it does for ChatGPT, Claude or Perplexity. Google's documentation says Search ignores the file. The two public datasets we could open and read, one covering nearly 300,000 domains and one covering 90 days of server logs on a single site, found no effect on AI citations or on crawler behavior. Both are vendor studies without independent review.

The file does have a job. It is a different job from the one it is often sold for. An llms.txt is a Markdown file, at the root of a site or under a path such as /docs/, that gives an AI agent a short map of where the useful content is. Its author, Jeremy Howard, describes it in the proposal as a way "to provide information to help agents use a website."

Helping an agent use a site is not the same as ranking in Search or being cited in an answer. Advice about llms.txt for SEO often merges the two.

What does Google say about llms.txt?

Google says Search does not use llms.txt, in classic results or in AI Overviews and AI Mode. The statement sits in its guide to generative AI features, in a list introduced as "a few things you can ignore for Google Search". The page was last updated on 2026-07-10. Two sentences from that list entry, verbatim:

Two details in that passage matter. "Ignores" is not "penalizes": Google says it is "completely fine" to keep the file for other services. And Google may still crawl and index the file like any other, which, in its words, "doesn't mean that the file is treated in a special way." Seeing Googlebot request your llms.txt is not evidence that it counts.

Why does Lighthouse audit llms.txt if Search ignores it?

Because Lighthouse is a different product asking a different question. Chrome's Lighthouse has an llms.txt audit among its agentic browsing audits. Its documentation gives the reason as agents: without the file, they "may spend more time crawling the site to understand its high-level structure." That page was last updated on 2026-05-05 and says nothing about rankings or citations.

The audit does not fail a site that has no file. Its documentation says a missing file is "marked as Not Applicable (N/A), as providing the file is optional at the moment." It flags only a server error when the file is requested.

Is llms.txt actually used by ChatGPT, Claude or Perplexity?

None of ChatGPT, Claude or Perplexity documents llms.txt as something its crawlers read to decide what to cite. We read the pages OpenAI, Anthropic and Perplexity publish for site owners. Each one tells you to manage its search crawler through robots.txt, and none lists llms.txt as an input.

The only llms.txt on the OpenAI and Perplexity pages is a link to the vendor's own documentation index, offered to agents reading those docs. That is the proposal's intended use, and it is not a statement about search.

Absence from documentation is not proof of absence. Some AI bots do request the file, as the log data in the next section shows, and any assistant that can open a URL can read an llms.txt a user points it to. What no vendor has published is that the file changes which sources get cited. We do not know what happens inside their retrieval systems; we know what they document.

Does Claude use llms.txt?

Anthropic publishes one, at platform.claude.com/llms.txt. It opens by describing itself as "an overview of the Anthropic API documentation and developer resources." That is a file for developers and their coding agents.

Anthropic's page for site owners is a separate document. It describes ClaudeBot, Claude-User and Claude-SearchBot and the robots.txt directives they honor, and it does not mention llms.txt. A company publishing the file for its own docs tells you nothing about whether its search crawler reads yours.

What does the measured evidence show?

The two public measurements we could verify both found no effect, and both come from companies that sell AI-visibility software. Neither is peer reviewed. We sell AI-visibility work too, so read this summary with the same suspicion.

SE Ranking's summary is direct: "There's no correlation between AI citations and LLMs.txt." It adds that removing the llms.txt variable from its prediction model "actually improved its accuracy." The largest sites were not ahead of anyone: among domains with 100,001 or more visits, 8.27% had the file.

OtterlyAI's result is about attention, not outcomes. By its own description the file was correctly implemented and directly accessible from the homepage, and AI bots still rarely asked for it.

Neither study proves the file can never matter. Both describe one period and one dataset, and both were published before the latest revision of the proposal page (modified August 10, 2026), which now describes the file in terms of agents. "No effect found" is the accurate reading, not "proven useless."

When is an llms.txt file worth publishing?

An llms.txt file is worth publishing when AI agents, not search engines, are the reader, which mostly means developer documentation. The proposal says so itself: "llms.txt files are used most heavily for software documentation, where coding agents follow them to find API references and tutorials." If you sell an API or an SDK, an agent that reads your docs correctly is worth something to the developer using it.

• Worth it: you publish API references, SDK guides or product documentation that developers load into coding agents.
• Worth it: your docs platform or CMS generates the file automatically, so it costs nothing and stays current.
• Not worth the meeting: a B2B marketing site with no documentation, where the goal is more citations in ChatGPT or AI Overviews.
• Not what it does: controlling crawler access. That is robots.txt.

One caveat applies even when the file is worth having. A hand-written llms.txt is a second copy of your positioning that nobody remembers to update. Generate it from the same source as the pages, or it will drift.

What should you check before spending time on llms.txt?

Before spending time on llms.txt, check that assistants can reach and read your pages, and whether they already name you. Both come before any new file, and both are free. Confirm the search crawlers are allowed in robots.txt, confirm the page still has text when JavaScript does not run, and measure whether assistants mention you today. If you publish developer documentation and still want the file, generate it from the docs build rather than maintaining it by hand. The steps below are in order.

1. Confirm the search crawlers are allowed. Open your robots.txt and look for OAI-SearchBot, Claude-SearchBot and PerplexityBot. OpenAI's documentation states that sites opted out of OAI-SearchBot "will not be shown in ChatGPT search answers."
2. Confirm the page has text without JavaScript. Request a key page without a browser and read what comes back. A page whose text only exists after scripts run can reach a crawler empty; we cover that failure in how to appear in Google AI Overviews.
3. Measure whether assistants mention you today. Ask the questions your buyers ask and record the answers. The method is in how to know if ChatGPT mentions your brand.
4. If you have developer docs, generate the file from the docs build. Follow the specification: an H1 with the name, a short summary, and sections of links to Markdown versions of the pages.
5. Judge it by your logs. Filter server logs for requests to /llms.txt by user agent. Do not credit the file with a citation change you cannot isolate from everything else you shipped that month.

What we have not measured

UPAI has no first-party figure on llms.txt. The check we have proposed compares two groups: domains cited in AI Overviews, and domains that rank for the same query but are not cited, counting how many in each serve the file. It has not been run, so we publish no number.

Our S&P 100 citation study measured which companies four assistants name, not which files those companies serve. It does not answer this question and we will not stretch it to.

What is known today: Google has said in writing that Search ignores the file, two vendor datasets found no effect, and three assistant vendors document robots.txt and say nothing about llms.txt as a citation input. For ChatGPT, Claude and Perplexity, the precise answer is that we do not know, and neither does anyone who has not published their logs.

Questions people ask next

These are the follow-ups we are asked most often once the answer above has landed. The same rule applies to each: where a figure appears, its source and year sit in the same sentence, and where we do not know something we say so rather than estimate it.

Is llms.txt mandatory?

No. It is a proposal, not a standard any search engine enforces. Chrome's Lighthouse documentation marks its llms.txt audit as Not Applicable when the file is missing, 'as providing the file is optional at the moment', and Google's Search documentation lists the file among things you can ignore.

Can an llms.txt file hurt my SEO?

Not in Google Search. Google's wording is that the file 'will neither harm nor help your site's visibility or rankings in Google Search, as Google Search ignores them.' The practical risk is different: a hand-written file that is never updated keeps describing old products and prices to any agent that does read it.

Does llms.txt replace robots.txt or a sitemap?

No. The proposal itself says robots.txt tells automated tools what access is acceptable, while llms.txt is read on demand when an agent needs information. It does not allow or block any crawler. Access for OAI-SearchBot, Claude-SearchBot and PerplexityBot is set in robots.txt, according to each vendor's own documentation.

Does Claude read llms.txt?

Anthropic publishes an llms.txt for its own developer documentation, which is the use the proposal describes. Its page for site owners describes ClaudeBot, Claude-User and Claude-SearchBot and the robots.txt rules they honor, and does not mention llms.txt. We found no Anthropic statement that the file affects which sites Claude cites.

How do I create and submit an llms.txt file?

Write a Markdown file named llms.txt and serve it at the root of the site or at a subpath such as /docs/llms.txt. The specification requires only an H1 with the site or project name; a short blockquote summary and H2 sections of links are optional. There is nowhere to submit it.