llms.txt Generator: Create, Upload and Check an llms.txt File (With Examples)

Someone told you every site needs an llms.txt now. Maybe it was your developer, maybe an SEO checklist, maybe a client. Then you opened the format and saw that it wants a hand-picked list of your best pages with a short summary on top. Which of your 300 URLs belong in it?

That's the job this llms.txt generator does in a few seconds. Enter your domain and it reads your sitemap, groups your pages by section, writes the file in the right format and hands it back ready to copy or download. You still decide what stays, but you start from a clean draft instead of a blank text file.

We'll also be straight about what the file does and doesn't do, because a lot of advice about llms.txt promises more than anyone has shown.

What an llms.txt file is

An llms.txt file is a short text file that lives at the root of your site, at yourdomain.com/llms.txt. It's written in markdown, the same simple format people use for notes and README files. Its job is to give AI tools a tidy map of your most useful pages. They don't have to dig through menus, cookie banners and scripts to find them.

The format is loose, but most files follow the same shape. A title line with your site's name comes first, then a one-line summary set off as a quote. After that come a few sections, each with a heading and a list of links, and each link can carry a short note on what the page covers.

Here's what a small one looks like.

# Your Company

> One sentence on what you sell and who it's for.

## Pages

- [Pricing](https://yourcompany.com/pricing): plans and what each one includes
- [Product tour](https://yourcompany.com/product): what the product does, screen by screen
- [About](https://yourcompany.com/about)

## Blog

- [How to pick a payroll provider](https://yourcompany.com/blog/how-to-pick-a-payroll-provider)
- [Payroll costs for a 10-person team](https://yourcompany.com/blog/payroll-costs-10-person-team)

Some sites also publish a second file, llms-full.txt, with the full text of their key pages in one place. It's optional, it can get very large, and you only need it if you want AI tools to read your docs or help pages in one go.

What the generator builds for you

The generator does the boring part of the job, the part where you'd copy URLs into a list for an hour.

1. It reads your sitemap. If your site doesn't have one, it reads the links on your homepage instead.
2. It takes your site name and your summary from the title and description your homepage already has.
3. It groups your pages by the first folder in their address, so blog posts land under Blog, help pages under Help, and so on.
4. It puts your top-level pages first, under Pages, since those are usually your homepage, pricing, product and about pages.
5. It keeps the file to 100 links and splits them fairly between sections.

That last rule came from our own site. The first version of the generator listed pages in sitemap order. When we ran it on rctrl.com, our 321 blog posts filled all 100 slots before a single product page made it in. A file like that tells an AI tool you're a blog. Now every section gets a fair share first, and your main pages always lead.

The result shows the finished file with Copy and Download buttons, plus a count of pages and sections. Each run also gets its own page, so you can send the result to whoever manages your site.

Will an llms.txt help your AI search traffic?

Probably not much, and we'd rather tell you that here than have you find out in three months. We don't know of any AI search engine that has confirmed it uses llms.txt to decide which pages to quote. When ChatGPT or Perplexity answers a question, it searches the web and reads pages directly. A file pointing to those pages doesn't seem to change which ones win.

The file can help AI tools and agents that are told to read a site, like a coding assistant reading your docs or an agent comparing vendors. A clean map saves them guesswork. That's a real but small benefit.

So our advice is simple. Spend two minutes on it, not two days. Generate the file, tidy it, upload it and move on to the work that actually brings visits, which is publishing pages that answer what your buyers search for.

When an llms.txt is worth more of your time

There are a few cases where we'd spend an extra hour on the file instead of two minutes.

If you sell software with public docs, an API or a help center, people now ask coding assistants and AI chat tools how to use your product. A clean llms.txt that points to your setup guides, API reference and common fixes gives those tools a shortcut. Some teams with big docs also add llms-full.txt for that reason.

If you run a large content site, a short list of your strongest guides is more useful than a dump of every post. Pick the 50 to 100 pages you'd hand a new customer, and leave the rest to your sitemap.

If you run a local service business or a small shop, keep it short and move on. Your service pages, pricing, location page and about page cover it.

What we found on 88 sites that rank on page one

We checked whether 88 sites had an llms.txt in October 2026. They were every site, other than forums and social networks, ranking in Google's top 10 for 12 everyday buyer searches. The searches covered software, online shops, home services, insurance and legal help, and agencies.

Twenty-nine of the 88 sites, about one in three, served a real llms.txt. Plenty of sites rank on page one without one, which tells you how much weight the file carries on its own.

Agencies led, which makes sense, since they sell this kind of work and often try it on their own sites first. Online shops barely bothered, with one file in 16 sites.

The files themselves were all over the place. The smallest had 11 links. The middle one had 59. The biggest had 989 links and weighed about 526 KB, which makes it a sitemap with extra steps rather than a short list of best pages. Eight files went past 100 links.

Most got the basics right. Twenty-two of the 29 included the one-line summary at the top, and all but two used sections to group their links.

Two small traps showed up too. Four sites answered /llms.txt with a normal web page, so a quick status check would say the file exists when it doesn't. Fourteen more sites refused our request for the file outright, the way bot shields often do. If you add a file, open it in a browser afterwards and make sure you see plain text.

Edit the draft before you upload it

A generated file is a strong first draft, and five minutes of editing makes it yours. Link titles come from the page addresses, so a page at /ai-visibility shows up as "Ai visibility". Fix the capital letters on the links that matter.

Then cut what doesn't belong. Sitemap pages, login and account pages, carts, thank-you pages and old campaign pages add noise. Keep your main product or service pages, your pricing, your best guides and the pages that answer the questions people ask before they buy.

Last, add a short note after your most important links, after a colon, like the pricing line in the example above. A note tells an AI tool what's on the page before it opens it. Check that your site name and summary read the way you'd describe yourself, since they come straight from your homepage's title and description.

Four mistakes to avoid

None of these break anything, but each one makes the file less useful to the tools that read it.

1. Listing everything. Eight of the 29 files we found went past 100 links. A file with hundreds of links is your sitemap in a new format, and it buries the pages that matter.
2. A slogan for a summary. "The future of work, reimagined" tells an AI tool nothing. Say what you sell and who it's for in one plain sentence.
3. Links that lead nowhere useful. Old campaign pages, redirects and pages your firewall blocks waste the reader's time, human or not.
4. Never updating it. Sites change their folders and products, and a file from last year can point at pages that no longer exist.

How to upload llms.txt on your platform

The file has to sit at the root of your domain, next to robots.txt. How you get it there depends on where your site runs.

Check that it's live

Open yourdomain.com/llms.txt in a browser. You should see plain text that starts with a line like "# Your Company". If you see your 404 page, a styled web page or a challenge screen, the file isn't being served the way AI tools need it.

It's also worth running our AI Crawler Checker once the file is up. An llms.txt does nothing if your firewall turns AI crawlers away before they reach it.

Keep it current

Run the generator again when you launch a new section, rename folders or retire a product line. It's free, and a stale file that points to dead pages is worse than no file at all. Once a quarter is plenty for most sites.

The file points to pages, and pages bring the traffic

An llms.txt is a table of contents. It can't make a thin site look deep. If your best pages are a homepage, a pricing page and a handful of old posts, the map is short because there isn't much to map.

What brings visits from Google and AI search is pages that answer what buyers type in, with real numbers and plain words. That's the work RankControl does. It finds the questions buyers in your market search. Then it writes and publishes up to 100 articles a month as native posts on your own site, whether that's WordPress, Webflow, Shopify or most other platforms. Auto-publish is off by default, so you can read each one first.

It tracks 50 of your buyer questions every week across six AI engines, Google AI Mode included. Your Google rankings and the visits AI answers send you sit right next to them. Traffic from Google and AI search shows up first.

Mentions inside AI answers come more slowly, mostly as other sites link to you and talk about you. Link Control helps with that part. It finds sites to pitch for each article and sends the emails from your own inbox after you approve them.

You can do this by hand. Write two strong articles a week, keep your llms.txt and sitemap current, pitch sites one by one, and check your questions in each AI engine every week. That's about 20 to 30 hours a month for a small team. Or RankControl can run it for $400 a month, with a 7-day free trial.