llms.txt Checker
Enter your website and the checker loads your llms.txt, reads it against the llmstxt.org format, checks that crawlers can fetch it, and tests every page it lists. You get a score out of 100, the exact lines to fix and how to fix them.
- Scores delivery, format, content and links
- Tests every listed page
- Paste or upload a file before you publish
Need a file to check? Build one with the llms.txt Generator, or see who can read your site with the AI Crawler Tester.
Four areas, one score
Each area has its own score, so you can see whether the problem is the file, the server or the links.
Delivery
Does /llms.txt load as plain text, without a redirect to another site, and without robots.txt keeping crawlers out?
Format
One # title, a > summary, ## sections and well-formed link lines, as the llmstxt.org proposal describes.
Content
A summary worth quoting, a note on each link, full https addresses and a focused list of pages.
Links
Every listed page is loaded to catch dead links, redirects and pages set to noindex.
Crawler access
robots.txt is read and tested for Googlebot and the main AI crawlers, so a Disallow rule does not hide the file.
llms-full.txt
The optional companion file is detected, so you know whether your site offers it.
How the 100 points are awarded
Checks that cannot be run, such as link tests on a pasted file, are left out and the score is worked out on the rest.
| Area | What is tested | Points |
|---|---|---|
| Delivery | The file loads, plain-text content type, no off-site redirect, not blocked by robots.txt, sensible size, no byte order mark | 25 |
| Format | One # title, a > summary, ## sections with links, every list item is a Markdown link | 30 |
| Content | Summary length, a note on each link, link count, full https addresses, no duplicates | 27 |
| Links | Listed pages load, no redirects, no noindex pages | 18 |
How to check your llms.txt
Three steps, about a minute.
Enter your site
Type your domain. The checker loads /llms.txt, robots.txt and the pages the file lists.
Read the score
Start with the errors, then the warnings. Each one names the line and says how to fix it.
Fix and check again
Edit the file, upload it, and run the check again until it is clean. Use the generator if you want to rebuild it.
What the checker usually finds
These are the faults that turn up most often when a site adds an llms.txt.
An HTML page instead of a file
Many servers answer a missing /llms.txt with their normal page and a 200 status. A model then reads a web page, not your file.
Dead or redirected links
Pages get renamed or removed. A listed link that returns 404 teaches a model nothing, and a redirect wastes a request.
No summary
The blockquote under the title is the first thing a model reads. Without it there is nothing to quote about who you are.
Relative addresses
Models fetch links on their own, so /about does not work. Each link needs the full https address.
The wrong content type
A file served as text/html or an unusual type may be skipped. Plain text or Markdown is safest.
Blocked by robots.txt
A broad Disallow rule can cover /llms.txt, so the file exists but the crawlers you wrote it for cannot fetch it.
What an llms.txt checker tests and why delivery matters
llms.txt is a proposed convention described at llmstxt.org: a Markdown file with a # title, a > summary and ## sections of links. Writing the file is half the job. The other half is making sure it actually reaches whoever fetches it, and that everything it points to works. That is what a checker is for.
The most common failure is not in the format at all. Many servers answer a request for a missing file with their normal page and a 200 status. Google calls this a soft 404: an error page served with a success code. A tool then reads a web page, not your file. A broad Disallow rule in robots.txt can also cover the file, so crawlers that obey it never fetch it.
Be clear about what a good score means. It shows the file is well formed, loads correctly and links to healthy pages. It does not mean search engines or AI products will use it. No major AI company has promised to, and Google's AI features documentation says no special files are needed to appear there.
When to check your llms.txt file
Check it when you publish it, and again whenever your site or server setup changes.
Right after you publish
Open the live address through the checker to confirm it loads as plain text, is not redirected somewhere else and has no format errors. Opening it in a browser tab is not enough.
Before you go live
Paste or upload the draft to check structure, summary and link formatting without publishing. If you leave the link test on, only the listed addresses are sent to the checking service.
After pages are renamed or removed
Links in the file go stale quietly. Re-run the check after a redesign, URL change or content clean-up to catch dead links, redirects and noindex pages.
After a security or CDN change
New firewall, bot or caching rules can change what is served at the root. Check that the file still loads and crawlers are not blocked from it.
On a new host or platform
Static hosts, WordPress hosting and CDN-based hosting handle root files differently. Confirm the file is served at the root of your domain with a sensible content type.
As part of an AI readiness review
Include the file in a wider review of crawl access, structured content and trust signals, remembering it is the least important of those.
Why an llms.txt check fails and how to fix it
Most failed checks come from the server, not from the writing.
The file returns an HTML page
What goes wrong: A missing file is answered with the site's normal page or custom error page and a 200 status. Tools then read HTML, not Markdown.
The fix: Upload a real llms.txt to the root, and make your server return a true 404 for missing files, as Google advises for missing pages.
Blocked by robots.txt
What goes wrong: A broad rule such as Disallow: / for a crawler group, or a rule that covers the file's path, stops crawlers that obey robots.txt from fetching it.
The fix: Test the file's path against the crawlers you care about, and add an Allow rule or narrow the block if needed.
Wrong content type
What goes wrong: A file served as HTML or a generic download may be skipped by a tool that expects text. Some servers guess types badly for unusual filenames.
The fix: Serve it as plain text with UTF-8 encoding, or as Markdown, and check the response headers after any server change.
Relative or broken links
What goes wrong: A tool fetching the file by itself cannot resolve /about. Links that return 404, or redirect to another page, give the reader nothing useful.
The fix: Use full https addresses, then fix or remove each link the checker reports. Link directly to the final URL, not a redirect.
Missing summary or notes
What goes wrong: A file with only a title and a list of bare links gives a reader nothing to understand or quote.
The fix: Add a one or two sentence summary under the title, and a short note after each link that says what the page covers.
Linking to noindex or private pages
What goes wrong: Pages with a noindex rule or a login wall contradict the purpose of a public guide to your best content.
The fix: Link only to public, indexable pages. If a listed page should not be indexed, remove it from the file.
llms.txt checking terms explained
- Delivery
- How the file reaches the requester: status code, content type, redirects and whether robots.txt allows access. A well-written file with poor delivery is of no use.
- Soft 404
- A missing page that returns a 200 status with error-style content. For llms.txt it means a crawler receives HTML where it expected your file.
- Content type
- The
Content-Typeheader the server sends. For a text file such as llms.txt, plain text or Markdown is safest; HTML or a download type is not. - Byte order mark (BOM)
- Invisible bytes at the start of some UTF-8 files. The llmstxt.org format allows an optional one, but plain files without it are the simplest to serve and read.
- Format check
- A test of the file's structure against the proposal: one title, an optional summary, sections with link lists and well-formed Markdown link items.
- Link test
- Loading each listed page to see whether it returns a good status, redirects or carries
noindex. It shows whether the file points to healthy pages. - Crawler access
- Whether robots.txt allows named crawlers, such as Googlebot or GPTBot, to fetch a given path. It applies to llms.txt as it does to any other file.
- llms-full.txt
- A companion file some sites publish with the full text of key pages in one document. It is a separate file from llms.txt and is not required.
Related tools, guides and services
If the check turns up problems, these tools help with the cause.
llms.txt Generator
Build or rebuild the file from your sitemap with a live preview.
ToolRobots.txt Tester
Test whether a rule in robots.txt keeps crawlers away from your llms.txt path.
ToolHTTP Header Viewer
Inspect the status code, content type and redirects served for /llms.txt.
ToolAI Crawler Tester
See which AI crawlers can reach your pages, including firewall blocks.
GuideFuture of SEO in Nepal
Puts llms.txt in context alongside AI search, crawler controls and trust signals.
ServiceAI SEO service
Get a plan for AI search visibility that goes beyond a single file.
Built by
Madhav Kushwaha
An llms.txt file is only useful if it loads properly and every link in it works, and those are the two things people forget to test. I built this checker to test the whole chain, from the server response to each listed page, and to point to the exact line that needs fixing. I run it on my own llms.txt too.
llms.txt Checker: Frequently Asked Questions
More free SEO tools
Make your site easy for AI tools to find and read.