Skip to content

Free tool · No signup · Part of the AI Visibility Check

llms.txt checker, validator and generator

llms.txt is a short text file some sites add to point AI tools at their best pages. Do you need one? Probably not yet: none of the big AI assistants has said it reads the file, and Google says its search ignores it. If you want one anyway, this page checks yours, tests a draft, or writes one for you. My guide explains what llms.txt is and whether your site needs one.

Check a live site's llms.txt

llms.txt is the smallest part of AI visibility. The full check tests whether AI tools can get into your site and read it, and gives you ten questions to see whether they recommend you.

llms.txt validator: paste a file

Paste the contents of an llms.txt. It's checked in your browser; nothing is sent.

llms.txt generator

Fill in the basics and list your most useful pages, one per line, as Section: Title | URL | note. The note is optional.

Last verified: October 2026

llms.txt examples, run through this validator

Three live files, each fetched and checked on October 29, 2026. They show the range: the reference, a large docs site, and a service business.

llmstxt.org: the reference file

The proposal's own llms.txt is short: an H1, a one-line blockquote, and a Docs section of three links with notes. It validates clean. If you're a small business, copy its shape, not its length. 

# llms.txt

> A proposal that those interested in providing LLM-friendly
> content add a /llms.txt file to their site. ...

## Docs

- [llms.txt proposal](https://llmstxt.org/index.md): The proposal for llms.txt

Cloudflare developer docs: an index of indexes

Cloudflare's file has nine H2 sections and 112 links, mostly to each product's own llms.txt. The validator flags one note: the first line after the H1 is plain prose, and the blockquote comes after it, so a strict reader misses the summary. Everything else parses.

connercrowe.com: a service business, with format notes

My own llms.txt has 16 sections and 152 links. It's valid, with warnings: H3 subheadings inside sections, and short paragraphs between the link lists. Both are readable, and both drift from the proposal. I left them in so this example is honest.

What the checker looks for

I fetch /llms.txt and /llms-full.txt from the root of your site. A file that comes back as an HTML page is the most common failure I see: the server answers every unknown path with its normal page, so there is no file at all. I report that as missing, with a note, not as a broken file.

The validator follows the format on llmstxt.org, in order:

  1. An H1 with the site or project name. The only required part. Missing it is an error, and so is a second H1.
  2. A blockquote summary right under it. Optional in the proposal, but it's the line a model reads first, so a missing one is a warning.
  3. Optional detail in any Markdown except headings.
  4. H2 sections of links. Each item is - [name](url), optionally followed by : notes. URLs with one level of parentheses, like a Wikipedia link, are fine. List items without a link and loose prose inside a section are warnings. A section called "Optional" marks links a reader can skip when it's short on room.

Here are those parts in a short example file, each labeled:

Example · fictional business example.com/llms.txt 
  1. # Riverside Plumbing

    The name

    Your business or site name, after one #. It's the only part the format requires, and the file gets exactly one, at the top.

  2. > Licensed plumber in Spokane, WA. Water heaters, drain cleaning and repiping for homes. Open Monday to Saturday, 7am to 6pm.

    The summary

    A line or two after a >, saying what you do. Optional, but it's the line a model reads first.

  3. Service area: Spokane and Spokane Valley. Prices are on each service page.

    Extra detail

    Any extra detail, in plain sentences or lists with no headings. You can skip it. It goes above the first section.

  4. ## Services
- [Water heater replacement](https://example.com/water-heaters/): tank and tankless, priced on the page
- [Drain cleaning](https://example.com/drains/): same-day in Spokane and Spokane Valley

## About
- [Service area and hours](https://example.com/contact/): cities served, phone, license number

    Sections of links

    A ## heading for each group of pages. Under it, one link per line: the page name in square brackets, the full address in parentheses, then an optional note after a colon. Full addresses are safer, because the file is often read away from your site.

  5. ## Optional
- [Blog](https://example.com/blog/)

    The Optional section

    A section called "Optional" holds links an AI tool can skip when it's short on room.

The parts of an llms.txt file, labeled, using a made-up plumber. The file is plain text saved at the root of the site, so it loads at yoursite.com/llms.txt.

Deeper headings get a warning, since the format only uses H2 for file lists. Relative links get an informational note, because the file is often read away from your site. Repeated problems fold into one line with a count, so a long file doesn't bury you.

How the generator builds the file

It writes the H1 from your name, the blockquote from your summary, an optional line of detail, then one H2 section per section name you use, with each page as a link and its note after a colon. The output is run back through the validator before you see it. Keep it to the ten to twenty pages that matter most; your XML sitemap and llms-full.txt are for the rest.

Google's guide to generative AI in Search puts both sides plainly: "It's completely fine if you decide to create and maintain LLMS.txt files" for other services, and doing so "will neither harm nor help" your visibility in Google Search, which ignores them. That's the honest case for the generator: ten minutes, no downside, and ready if an assistant starts reading the file.

Where this fits

llms.txt is one line in the "can AI read you" part of the AI Visibility Check, and it carries the least weight of anything there on purpose. The check also tests crawler access, server rendering and schema, and gives you ten prompts to find out whether ChatGPT, Google and Perplexity name your business at all. The robots.txt checker for AI bots covers the access part on its own, bot by bot, and the AI crawler directory explains each bot. If you'd rather have me ask the AI tools for you, that's the free AI Answer Snapshot.

Sources

llms.txt checker questions

What is llms-full.txt?
A companion file many sites publish next to llms.txt, with the full text of their pages inlined so a model can read everything in one request. The llmstxt.org proposal doesn't define it; it grew out of practice. The checker confirms it exists and isn't a web page.

What makes an llms.txt file valid?
One H1 with the site name at the top. That is the only required part. After it, a blockquote summary, optional prose, then H2 sections that each hold a list of Markdown links, optionally followed by a colon and a note.

Where does llms.txt go?
At the root of your site, so it loads at yourdomain.com/llms.txt, served as plain text. If that address shows your normal web page, the server is answering every unknown path with HTML, and the checker will tell you so.

Is this llms.txt checker free?
Yes. No signup and no email. Pasted files and the generator run in your browser and are never sent anywhere.