About Validating llms.txt
What there is to validate
The llms.txt format is short: an H1 with the site name, an optional blockquote summary, then H2 sections whose bodies are markdown link lists written as name, URL and description.
Because so little is strictly required, a file can be perfectly valid and still be worth almost nothing - one heading and a bare link passes every structural test and tells an assistant nothing.
So this tool separates the two questions. Errors are what breaks the format; suggestions are what stops the file being useful.
What breaks the format
- No H1 heading, which is the one genuinely required line.
- More than one H1, leaving the site name ambiguous.
- List items that are not markdown links.
- Links with no text between the brackets.
- Relative paths, which are ambiguous once the file is read away from your site.
What makes it worth having
- A summary line, so a reader knows what the site is before reading any links.
- Section headings, so each link belongs to a part of the site.
- A description on every link that says what is on the page rather than repeating its name.
- Restraint. A file with hundreds of links is a sitemap, and curation was the point.
The descriptions do most of the work here. They are what makes a page selectable for a particular question instead of merely listed.
What this tool checks
Enter your domain and it fetches /llms.txt, reads the file, and reports the structural errors and the quality suggestions separately, along with a summary of what the file contains.
It also catches the soft 404 case, where a server answers 200 with its normal page instead of a file - which looks like success in a browser and is the most common false positive.
It reads the file as published rather than fetching each link, so a listed page that has since moved will not be flagged here.
Where to go next
If there is no file yet, our
llms.txt generator will build one from your site details. The
llms.txt checker answers the simpler question of whether a file exists at all, and the
AI crawler checker confirms the pages it points at can actually be fetched.