UPDATE
“FWIW no AI system currently uses llms.txt.” — John Mueller confirmed that mainstream AI services like OpenAI, Google Gemini, and Anthropic Claude aren’t even checking for the file
In SEO and AI, there’s always something new to tinker with. The latest buzz? A little file called llms.txt.
You might’ve seen devs or marketers talking about adding this file to your site to help large language models (LLMs) like ChatGPT or Claude “understand” your content better.
Sounds promising, right?
Well… maybe. Let’s unpack what llms.txt actually is, who’s using it, and whether it matters for your SEO strategy.
What is llms.txt?
Think of llms.txt as a Markdown version of a sitemap — but made specifically for large language models.
It’s a simple text file, hosted at the root of your domain (like https://yourdomain.com/llms.txt), that lists out your site’s most useful content: API docs, return policies, product data, sizing guides, and so on.
The idea is to give AI models a curated guide to your best structured content, so they don’t have to guess what matters.
Here’s an example:
# llms.txt
## Products
- /catalogue.md
A structured list of categories, products and SKUs.
- /size-guide.md
Size charts for different types of clothing – shirts, pants, shoes, etc.
## Policies
- /shipping.md
Details on delivery options, timeframes and costs.
- /returns.md
Return policy with info on time limits, conditions and how to send things back.
## Help
- /faq.md
Answers to common questions about orders, payments and accounts.
- /contact.md
How to get in touch with our customer support team.But does anyone actually use it?
A few companies have jumped on board. You’ll find llms.txt files on sites like:
- Cloudflare (performance and security docs)
- Anthropic (Claude’s parent company)
- Tinybird (real-time data APIs)
- Mintlify (developer docs platform)
There’s even a public directory here: directory.llmstxt.cloud
But here’s the kicker: no major LLM provider actually supports it — not officially.
- OpenAI (ChatGPT): Respects robots.txt, but no mention of llms.txt.
- Google (Gemini/Bard): Uses Google-Extended in robots.txt to manage crawling for AI, but nothing on llms.txt.
- Meta (LLaMA): No known crawling protocol or use of llms.txt.
- Anthropic (Claude): Has a llms.txt file… but doesn’t confirm whether Claude uses it.
So right now, it’s basically a hopeful suggestion — like writing “please send me traffic” on your homepage.
Is it useful for SEO?
Let’s be real: not today.
There’s no data showing that llms.txt helps your visibility in AI search or chatbots. No evidence that LLMs parse or prioritise it. No commitment from any model provider that they’ll ever support it.
And as Google’s John Mueller said on Reddit:
So… useful? Not yet.
But harmful? Also no.
Should you add one anyway?
If you’ve got strong structured content — product data, developer docs, policy pages — then sure, why not?
- It’s easy to create.
- It won’t hurt anything.
- And if llms.txt does get adopted later, you’ll be ahead of the curve.
You can even use a free generator to make one like that one:
Just don’t expect a traffic spike or better AI citations because of it.
The real takeaway for SEOs
Focus your energy where it counts today:
- robots.txt — Make sure it’s not blocking key pages.
- sitemap.xml — Help search engines find and index your content.
- structured data — Schema.org markup still matters.
- content quality — That’s what LLMs are really looking for.
llms.txt might become useful one day. But until LLMs start saying “yes please” instead of “never heard of her”, it’s more curiosity than necessity.
Final thoughts
- llms.txt is a proposed file for helping LLMs find your best content.
- No major AI model currently uses it.
- It’s easy to set up, but there’s no real benefit (yet).
- Worth doing if you’re curious, but don’t lose sleep over it.
If anything changes — say, OpenAI or Google starts crawling llms.txt — you’ll hear it here first.
Got questions or spotted a use case in the wild? Ping me on LinkedIn or send a carrier pigeon 🕊️
