No — llms.txt barely gets you into AI answers. According to Otterly's 90-day monitoring of a live site, just 84 of 62,100 AI-crawler visits touched it (0.1%), and no major provider confirms using it. What actually opens the door to ChatGPT and Perplexity is crawler access, a complete sitemap, and clean server-side rendering.

llms.txt is a plain-text file placed at the root of your domain — a machine-readable summary of who you are, where you are, and what you do, written for large language models instead of for people. AEO (Answer Engine Optimization, sometimes written GEO) is the craft of getting cited by AI assistants like ChatGPT, Perplexity, Gemini and Claude. llms.txt is the single most-recommended AEO tactic of 2026 — and, as the measured data below shows, also the most overrated.

Watch the walkthrough

What exactly is llms.txt, and why does every AEO guide recommend it?

llms.txt is a plain-text summary of your business — who you are, where you are, what you do — written for the language models rather than for human readers. The idea is genuinely appealing: instead of making a model guess from your marketing pages, you hand it one tidy, machine-readable description. That is why nearly every 2026 AEO checklist tells you to add one. The intention is good; the evidence is where it falls apart.

What does the data actually show about llms.txt?

According to Otterly's 90-day monitoring of one live site, out of 62,100 AI-crawler visits, just 84 requested the llms.txt file — one tenth of one percent. And it is not a single-site fluke: an analysis of roughly 300,000 domains published by Lumentir found no relationship between having llms.txt and how often a site is cited by the big models. The file almost every guide sells is, in the measured reality, barely fetched.

Does Google use llms.txt — and where is it genuinely useful?

No. Google's John Mueller has said plainly that no AI system currently uses llms.txt, and compared it to the old keywords meta tag — a field site owners fill in that engines long ago learned to ignore. No major provider — not OpenAI, not Anthropic, not Google — has confirmed reading it to build their answers. The one place llms.txt genuinely earns its keep is developer documentation, where coding assistants pull API references — the technical docs behind tools like Stripe or Vercel. For a small business trying to get into ChatGPT, it is close to noise.

Which AI crawlers must your robots.txt let in?

Your robots.txt is the first file a crawler reads, and it decides who gets in at all. It should explicitly allow the named AI crawlers — GPTBot and OAI-SearchBot from OpenAI, ClaudeBot from Anthropic, PerplexityBot from Perplexity, Google-Extended from Google. This matters more than any summary file: OpenAI's own documentation is blunt that if you block OAI-SearchBot, you are absent from ChatGPT's search answers — not ranked low, absent. Most small sites block these bots by accident, with a single line copied years ago. Opening that door is free, and it is the single highest-value move here.

What does your sitemap.xml actually do?

Your sitemap.xml is simply a list of every page you want crawled, with the date each one last changed. It is how a crawler finds pages that are not linked from your front page, and how it learns which pages are fresh. A thin or missing sitemap leaves the crawler with only whatever it stumbles onto from your home page; a complete one hands it the whole map at once. Most platforms generate it automatically — your job is to confirm it exists, that it is listed in your robots.txt, and that it holds your important pages. Unlike llms.txt, every engine uses this one.

Why does server-side rendering decide whether crawlers see your words?

Many modern sites build their content with JavaScript. The problem is that a lot of AI crawlers read the raw HTML your server sends first — and if your words only appear after the JavaScript runs, the crawler can land on a nearly empty page. The fix is server-side rendering: make sure the text of your page is in the HTML off the server, before any script runs. You can check it in seconds — open the page source, and if you can see your real paragraphs, so can the crawler.

Does any of this work differently in Polish or German?

The plumbing does not — robots.txt, sitemap.xml and server-side rendering are language-blind, and a crawler reaches a Polish or German page through exactly the same three doors. What changes is the answer waiting on the other side. Once the door is open, the engine matches a native-language question to native-language words, not to a translation, so the citation-winning move for a non-English business is a clear, direct answer written in the customer's own language. Get the plumbing right once — it is free and mostly automatic — then spend your real effort there. And llms.txt? Keep it if you already have one, since it costs nothing; but no engine has confirmed using it, so put no real effort into it.

Frequently asked questions

Does adding an llms.txt file help me get cited by ChatGPT?

Barely. According to Otterly's 90-day monitoring, only 84 of 62,100 AI-crawler visits touched llms.txt — 0.1% — and a separate look at roughly 300,000 domains found no link between having the file and being cited. No major provider confirms using it. Add it if you like, since it costs nothing, but do not expect it to move you into AI answers.

Do OpenAI, Anthropic or Google actually read llms.txt?

None has confirmed reading it to build answers. Google's John Mueller has stated plainly that no AI system currently uses llms.txt and likened it to the old keywords meta tag. Otterly's research reaches the same conclusion — no major LLM provider has publicly committed to using it at scale.

Should I delete my llms.txt file, then?

No need — it costs nothing to keep, and it is genuinely useful in one place: developer documentation, where coding assistants pull API references (the docs behind tools like Stripe or Vercel). Just do not rely on it as a ChatGPT or Google trick, and do not let anyone sell it to you as one.

What is the single highest-value AEO fix instead?

Crawler access, set in your robots.txt. It is the first file a crawler reads and it decides whether AI bots get in at all. OpenAI's own documentation says that if you block OAI-SearchBot, you are simply absent from ChatGPT's search answers. Opening that door is free — and most small sites are blocking these bots by accident.

How do I check which AI crawlers can reach my own site?

Open your own robots.txt and confirm the named AI crawlers — GPTBot, OAI-SearchBot, ClaudeBot, PerplexityBot, Google-Extended — are allowed, then view your page source to confirm your real text is in the server HTML. Our open-source tracker, aeo-platform, checks crawler access for you automatically, and it is free on npm.

Want to know which AI crawlers can actually reach your site?

We run the full check on your site — which engines can reach you, which pages they see, and which queries you are missing — and the first AI-visibility audit is free. If you would rather stop guessing about your robots.txt, sitemap and rendering, request a free audit at webappski.com.