What is llms.txt?
llms.txt is a plain-text file in Markdown format at yourdomain.com/llms.txt. In a few lines it says what your website is about and links to your most important pages. It's written for large language models like ChatGPT, Claude or Perplexity that read your site while answering a question and need to find the right page fast.
Jeremy Howard proposed the format on llmstxt.org on September 3, 2024. It's an open proposal, not an official standard like robots.txt.
How the generator builds your file
- 1
Claire opens your homepage and reads your robots.txt. It usually says where your sitemap is; otherwise we look in the usual places. We read a sitemap index one level deep.
- 2
From the sitemap we pick up to 60 pages of your domain: shallow pages first, then the newest deeper ones. Files, legal pages, account and cart pages and anything your robots.txt blocks stay out.
- 3
For the first 30 pages we fetch title and meta description from the page itself. We strip your brand name from the title so that not every line ends the same way.
- 4
We group pages by area, such as blog, guides or products. Anything beyond the limits goes into the “Optional” section at the end.
We store neither your website nor the result.
llms.txt example: how the file is structured
Here is an example for a made-up coffee roaster, with every part llmstxt.org describes:
# Example Roasters1 > Small-batch specialty coffee, roasted in Portland and shipped across the US.2 ## Main pages3 - [Shop](https://roasters.example/shop): Single origins, blends and subscriptions- [Contact](https://roasters.example/contact) ## Guides - [Brewing guide](https://roasters.example/guides/brewing): Pour-over, French press and espresso step by step- [Light vs. dark roast](https://roasters.example/guides/roast-levels) ## Optional4 - [About us](https://roasters.example/about)
- 1The first line with “#” and the name is the only required part.
- 2The line starting with “>” sums up in one sentence what you offer and for whom.
- 3Each section starting with “##” holds a list of links. A short note may follow the colon.
- 4According to llmstxt.org, “Optional” is the convention for secondary information a model may skip when it's short on space.
Where to put your llms.txt file
The file has to be reachable at yourdomain.com/llms.txt. Some platforms now create it for you. Here is how it works on the common ones:
WordPress
Easiest via your SEO plugin. Yoast SEO: Settings → Site features → turn on “LLMS.txt”. Rank Math: turn on the “LLMS Txt” module in the dashboard, then edit it under General Settings → Edit llms.txt. All in One SEO: Sitemaps → “LLMs.txt” tab (on by default). If you want to use exactly this suggestion, upload the file via FTP or your host's file manager to the root directory (next to wp-config.php) and turn off the plugin feature.
Shopify
Since May 2026 Shopify creates an /llms.txt automatically. You can customize it in the theme code (Online Store → Themes → Edit code) with the template templates/llms.txt.liquid. Paste the suggestion there.
Squarespace
Version 7.1 only: SEO settings → “LLMS.txt” tab → paste the text → save. The feature is off by default.
Wix
Wix creates the llms.txt automatically once your site has a Premium plan, a custom domain and active indexing: SEO & GEO → Tools and settings → llms.txt. If you edit the file there, Wix stops updating it automatically.
Webflow
Site settings → SEO → LLMs.txt → upload file (UTF-8, under 100 KB). Webflow only serves it on your custom domain, not on the staging domain.
Shopware 6
From version 6.7.12, according to the release notes, Shopware can serve an /llms.txt per sales channel itself. In older versions, put the file in the public/ folder of your installation.
TYPO3
Via a static route: in the site configuration (config/sites/…/config.yaml or the Sites backend module) add a route “llms.txt” of type staticText and paste the content.
Jimdo
We haven't found an official way to place a file at /llms.txt on Jimdo. According to its help center, Jimdo doesn't allow your own files in the root directory.
Other platform / your own server
Save the file as llms.txt (UTF-8) in your website's root directory so it's reachable at yourdomain.com/llms.txt. With Next.js, Astro or Vite it goes into the public/ folder.
As of October 2026
What llms.txt does and doesn't do
llms.txt is an offer to AI systems, not a ranking switch. It doesn't decide whether ChatGPT recommends you. That comes down to content that answers your customers' questions and to mentions on sites the models trust.
More important than llms.txt is that AI crawlers are allowed to read your website at all, and able to. If your robots.txt blocks them, or your text only appears after JavaScript runs, the best llms.txt won't help. That's exactly what the AI crawler checker tests.
Frequently asked questions
Is the llms.txt generator free?
Yes. You don't need an account or an email address. To keep it fast for everyone, you can run it up to eight times every ten minutes.
Which pages does the generator include?
Up to 60 pages of your domain from the sitemap: shallow pages first, then the newest deeper ones. Files, legal pages like imprint and privacy policy, account and cart pages and anything your robots.txt blocks stay out. We fetch title and description for the first 30 pages; for the rest we derive the title from the URL. You can change everything in the editor.
Do I also need an llms-full.txt?
The description on llmstxt.org only covers the llms.txt. Some plugins additionally create an llms-full.txt with the full text of all pages. It isn't required.
Does the llms.txt replace the robots.txt?
No. The robots.txt controls who may fetch your site. The llms.txt only helps those who may to find their way. If your robots.txt blocks AI crawlers, the best llms.txt won't help.
How often should I update the file?
Whenever important pages are added or removed. Shopify, Wix and the WordPress plugins keep their automatically generated version up to date. An uploaded file is yours to maintain.
Do you store my data?
No. We fetch your website, build the suggestion and send it to you. Neither the address nor the result is stored. To limit requests we briefly count in memory, without storing your IP address.