Someone has told you that your website needs an llms.txt file to show up in ChatGPT. It may have been a developer, or part of an agency pitch. The short answer is that Google Search does not use one, that an Ahrefs study found 97% of the files it checked were not requested even once in May 2026, and that nothing we could verify shows it earns you citations in other AI tools. If your real question is why your business is missing from AI answers, the work starts elsewhere: crawler access, indexed pages that answer the question, and checking whether you are named. That is where our GEO service starts too.
What llms.txt is
An llms.txt file is a plain text file, written in Markdown, that usually sits at the root of a website (for example, yoursite.com/llms.txt). Ahrefs describes it as "a single index file, written in markdown, placed at a site's root." The idea is to give AI tools a short summary of your business and key pages.
Jeremy Howard proposed it in September 2024, and the proposal's own site still calls it a proposal rather than a standard. That site says the files are used most heavily for software documentation, where coding agents follow them to find API references. If you publish developer docs, that is the case for one. The file blocks and allows nothing, and it only matters if some system chooses to read it.
What Google says about llms.txt
Google says you do not need one for Google Search, including AI Overviews and AI Mode. Its official guide to generative AI features lists llms.txt files under things you can ignore, and says Google Search itself does not use them. The guide adds that it is "completely fine" to create and maintain one for other services that use these files, and that doing so will neither harm nor help your visibility or rankings in Google Search.
Google's Search Central changelog records this clarification on June 15, 2026, and the guide page shows a last-updated date of July 10, 2026. It says nothing about ChatGPT, Claude or Perplexity.
What the Ahrefs study found
Ahrefs found that 97% of the llms.txt files it verified received no requests at all in May 2026. The study was published on June 15, 2026 by Louise Linehan. Ahrefs looked at 137,210 domains in Ahrefs Web Analytics that received traffic in May 2026. About 38,000 of them, 28%, had a valid llms.txt file (a real Markdown file at the site root, not an error page; Ahrefs did not check files against the specification). Of those files, 97% saw no requests for them in May, from bots or from people. Read the scope before you repeat the number.
- The sample is Ahrefs Web Analytics customers, who Ahrefs says skew more technical and SEO-aware than the web at large, so it treats the 28% adoption figure as an upper bound.
- It covers one month and counts requests to the file. It is not a test of whether a site gets cited in AI answers, so it cannot prove the files have no effect.
- Ahrefs' own conclusion is that for showing up in ChatGPT, Perplexity or AI Overviews the file is "largely decoration."
What is not known about other AI systems
None of the AI companies we checked tells site owners to add llms.txt, and the crawler pages OpenAI and Anthropic publish for site owners talk about robots.txt instead. Ahrefs did log a few fetches: OAI-SearchBot, PerplexityBot and Claude's search crawler made only a couple of hundred requests for llms.txt files across thousands of sites in May 2026, and Ahrefs notes that a fetch does not show the file was read or used. We would rather say that plainly than guess.
- OpenAI: its crawler page (checked October 2, 2026; the page shows no date) recommends allowing OAI-SearchBot in your robots.txt file so your site can appear in ChatGPT search results. It does not mention llms.txt as something site owners should add. The only mention is a link to OpenAI's own documentation index.
- Anthropic: its help article on its crawlers, dated April 7, 2026, says its bots honor robots.txt directives. It does not mention llms.txt.
- Perplexity and Microsoft Copilot: we did not open their publisher pages, so we make no claim either way.
That is not proof that no system reads the file. It means nobody has published a rule you can rely on, and an unproven file should not be sold as a way into ChatGPT.
Where this site stands
GrossiWeb's own site serves an llms.txt file at grossiweb.com/llms.txt. It was cheap to add, and Google says a file like this neither helps nor harms a site in Google Search, but the evidence above does not show that it brings a benefit, and we are not claiming it is the reason anyone sees us in an answer. If you keep one, keep it accurate: Ahrefs points out that a stale or tampered file misleads any agent that reads it.
If someone quotes you real money for llms.txt, or sells it as a way into ChatGPT, ask what evidence supports that. The same goes for "AI schema" packages.
What about AI schema and FAQ markup packages
Google says structured data is not required for generative AI search, and there is no special schema.org markup you need to add. It still calls it a good idea to keep using structured data as part of your overall SEO, because it helps with rich results eligibility.
One specific case: Google's changelog says the FAQ rich result is no longer shown in Google Search. Do not pay for FAQ markup expecting a rich result. There is no FAQ rich result left for that markup to earn, so do not count it as a growth lever.
What to do with the same hour instead
Spend it on things that are documented to matter: letting crawlers fetch your pages, writing pages worth citing, and checking whether you appear. Google's guide says optimizing for generative AI search is optimizing for the search experience, and so still SEO, which makes SEO fundamentals the base.
- Check that crawlers can reach you. Open yoursite.com/robots.txt and confirm you are not blocking OAI-SearchBot, because OpenAI says sites that opt out will not be shown in ChatGPT search answers. Ask your host whether bot protection blocks AI crawlers too.
- Confirm your key pages are indexed in Google. The guide says the way Google finds and processes your pages remains the core of how its AI systems access your data.
- Write the page only you can write. Google's guide gives "7 Tips for First-Time Homebuyers" as commodity content that adds little, and asks for first-hand content instead. Our post on how GEO relates to SEO covers structure.
- Measure before you change anything. Ask your buyers' questions in ChatGPT and Google AI Mode and write down whether you are named. Our post on measuring SEO and AI answers before the click explains the method, and analytics and reporting covers the reporting side.
If the answer to that last step is that you are absent, our page on why your business is not in AI answers lists the most common reasons AI tools skip a business. Work through it before you pay for any file, markup or package.
What to do this week
- Open yoursite.com/llms.txt and yoursite.com/robots.txt in a browser and note what each contains.
- If an agency or developer proposed llms.txt, ask in writing what evidence shows it improves visibility.
- Pick five questions your customers ask and run them through ChatGPT and Google AI Mode. Record who is named.
- Rewrite the first two sentences of one important page so they answer the customer's question directly.