llms.txt Checker
Check that your llms.txt exists, follows the format and can be reached by AI crawlers.
| Part | Content | Links |
|---|
What is llms.txt?
llms.txt is a plain Markdown file that sits at the root of a website and gives AI models a short map of it. It names the site, says in one line what the site is, and links to the pages that explain it best.
The idea comes from a simple problem. A language model can only read so much at once, and a normal web page is mostly navigation, scripts and layout. llms.txt hands the model a clean list of the pages that matter, in a format it reads easily.
What goes in the file
Four things, in this order: the name of the site, a short summary, any background a reader needs first, and groups of links. Each link can carry a note that says what the page covers. That is the whole format. There is no code in it and no settings, only text a person can read too.
Where the idea came from
The format is published at llmstxt.org as an open proposal. No standards body stands behind it and nobody has to support it, which is why the question of who reads the file matters more than the format itself.
Who uses it today
Mostly documentation. Software companies publish an llms.txt so that coding tools and agents can find the right page of the docs without crawling the whole site. People also paste the file into a chat to give a model the context in one go. Whether the big assistants read it on their own is a different matter, covered further down this page.
What this llms.txt checker checks
That the file exists. It has to open at /llms.txt on your host and come back as text. A redirect to the homepage or an error page does not count.
That the format is valid. One title, a summary, sections, and links written the way the format expects. Every problem is listed with its line number. That's the llms.txt validator part.
That AI crawlers can reach it. The checker reads your robots.txt and tells you if any search or AI crawler is told to stay away from /llms.txt. A file nobody may fetch does nothing.
Whether llms-full.txt is there. It is optional, and the result says if you have one.
You also get an outline of the file: the title, the summary, and each section with the number of links in it. Once the file is fetched, the checking happens in your browser. A file you paste or edit never leaves the page.
The llms.txt format, with an example
The file is ordinary Markdown with a fixed order. Only the title is required.
# Acme Analytics
> Acme Analytics is a web analytics tool for online shops. It tracks visits, orders and revenue without cookies.
Pricing is per site. All plans include the API.
## Docs
- [Getting started](https://acme.example/docs/start): install the script and see the first data
- [API reference](https://acme.example/docs/api): every endpoint with examples
## Product
- [Pricing](https://acme.example/pricing): plans and what each one includes
## Optional
- [Changelog](https://acme.example/changelog): release historyThe title
The first line is a single # followed by the name of the site or product. There can be only one.
The summary
A line starting with > right under the title. It says what the site is in a sentence or two. A model that reads nothing else should still know what you do.
Free text
Any paragraphs or plain lists after the summary. Use them for facts a model needs before it follows a link. No headings here.
Sections with links
Each ## heading starts a section, and a section is a list of links. One link per line, written as - [Title](https://full-url): short note. The note is optional and worth writing, because it tells the model what it will find before it spends anything on fetching the page.
The Optional section
A section named Optional has a special meaning. A tool that is short on space may skip it. Put the nice-to-have pages there.
llms.txt vs llms-full.txt
The two files answer the same need at different sizes.
llms.txt is the index. It is short, and a model that wants more has to follow the links. llms-full.txt is the content itself: the full text of your key pages joined into one Markdown file, so a tool can load everything in a single request.
The long file grew up around the format as a convention, mostly on documentation sites, where reading the whole manual at once is useful. A marketing site with a handful of pages rarely needs it.
If you publish both, keep them in step. A long file that describes last year's product is worse than no file. The checker tells you whether /llms-full.txt exists on your host and treats a missing one as normal.
Does llms.txt actually work?
The honest answer has two parts.
No major AI assistant has confirmed that it reads llms.txt when it answers a question. If you add the file and expect ChatGPT or Gemini to start citing you because of it, nothing supports that.
The file is used in other places. Coding tools read it to learn a library's documentation. Agents open it when someone points them at a site. People paste it into a chat to give a model the context in one go. For a product with documentation, that is real use today.
So treat it as a small, cheap addition. It takes an hour to write and costs nothing to keep. It does not replace the work that decides whether assistants name you, which is being worth citing and being mentioned on the sites they already trust.
llms.txt vs robots.txt vs sitemap.xml
The three files sit next to each other at the root of a site and do different jobs.
| File | What it does | Who reads it | Can it block access? |
|---|---|---|---|
robots.txt | Tells crawlers which URLs they may fetch | Search and AI crawlers | Yes, for crawlers that follow it |
sitemap.xml | Lists every URL you want found | Search engines | No |
llms.txt | Points AI models at the few pages that explain the site | Tools and agents that choose to | No |
The common mistake is to expect llms.txt to control anything. It can't keep a crawler out and it can't let one in. If robots.txt blocks a crawler from your pages, listing those pages in llms.txt changes nothing. Check what your file tells each crawler with the Robots.txt Tester.
How to create an llms.txt file
- Pick the pages. Choose the ones that explain what you do: the product page, pricing, the main docs, the page that answers the question you get most. Leave out everything else.
- Write the title and the summary. The name on the first line, then one or two sentences under
>that a stranger would understand. - Group the links. Give each group a
##heading and list the pages with full URLs and a short note each. - Save it as
llms.txtat the root. It has to open atyoursite.com/llms.txtas plain text. - Check it. Paste the draft into the checker before you publish, then run the live address once it is up.
Write it by hand if you can. A file generated from the sitemap lists everything, and a list of everything is the opposite of what the format is for.
llms.txt best practices
Write a summary that explains, not a slogan. "The fastest way to grow" tells a model nothing. "A booking tool for dental clinics" does.
Keep the list short. Pick the pages you would send to someone who has never heard of you. If you can't decide between two pages, the weaker one goes under Optional.
Describe every link. The note after the colon is what a model uses to decide whether the page is worth opening.
Use full addresses. Every link starts with https:// and points at the final URL, not at a redirect.
Link to pages that read well as text. A page that shows its content without scripts or a login is the one a model can use.
Update it when the site changes. Dead links and old prices in this file do more harm than a missing file.
Leave it open in robots.txt. The file is only useful to crawlers that are allowed to fetch it.
Common llms.txt mistakes
The file is not at the root. /docs/llms.txt or /static/llms.txt is not where tools look.
The address returns the homepage. Many sites answer any unknown URL with a page instead of an error. It looks like the file exists and it doesn't.
No title, or two of them. The file starts with one # line. Everything below it uses ##.
Relative links. /pricing means nothing to a tool that got the file from somewhere else. Use the full address.
Every page on the site. A long list defeats the purpose. The value is in what you leave out.
robots.txt blocks the file. A broad Disallow can cover /llms.txt without anyone meaning it to.
Links to pages a model can't read. If the page needs a login or builds its content with scripts, the link leads to nothing useful.
Frequently asked questions
What is llms.txt?
llms.txt is a plain Markdown file at the root of a website that gives AI models a short map of the site. It names the site, says in one line what it is, and links to the pages worth reading. It is a proposed format, not an official web standard.
Where does the llms.txt file go?
At the root of the host, so it opens at yoursite.com/llms.txt. A file in a subfolder is not found by tools that look for it. Each subdomain needs its own.
How do I check my llms.txt?
Enter your site in the checker above. It fetches the file and tells you if it exists, if the format is valid, and whether your robots.txt keeps AI crawlers away from it. To check a draft, paste it instead.
Do ChatGPT, Claude and Gemini use llms.txt?
No major assistant has confirmed that it reads llms.txt when it answers a question. The file is used today by coding tools, by agents that are pointed at it, and by people who paste it into a chat. Treat it as a low-cost addition, not as a ranking lever.
Is llms.txt the same as robots.txt?
No. robots.txt tells crawlers where they may go. llms.txt lists the pages you'd like AI models to read. It doesn't allow or block anything, so a page you want kept out still needs a rule in robots.txt.
What is llms-full.txt?
A companion file that holds the full text of your key pages in one Markdown document. llms.txt is the short index with links, llms-full.txt is the whole content. It is optional.
Does llms.txt help SEO?
Not in Google. Search rankings don't use it. Its only possible effect is on tools and agents that choose to read it, and the file doesn't replace anything you already do for search.
How long should an llms.txt file be?
Short. It is an index, so list the pages that explain your product best and leave the rest out. If a section is nice to have but not needed, put it under a heading called Optional, which tools may skip.
Do I need an llms.txt generator?
No. The file is short enough to write by hand, and a hand-picked list beats a generated one. Most generators copy the sitemap, which lists every page. If you use one, cut the result down to the pages that matter and run it through the checker.
How do I add llms.txt to a WordPress site?
Several SEO plugins can create the file from your settings. Without a plugin, upload a file named llms.txt to the root folder of the site, next to robots.txt. Then run the checker to confirm it opens at /llms.txt.
A clean llms.txt tells a model where to look. Whether the assistants name your brand when someone asks is a separate question, and the AI Visibility check answers it.