Checks whether your llms.txt exists, is actually
markdown rather than an HTML page returned with a 200, conforms to the spec, and
can be found. Free, no signup, nothing stored.
Checks /llms.txt, /llms-full.txt and your
robots.txt. Results cached five minutes.
Ahrefs found that across 137,210 domains, 97% of llms.txt files received zero requests in May 2026. If your file is broken, this tool will tell you — but the honest context is that most llms.txt files currently appear to go unread. Fix yours because you want it correct, not because it is proven to do anything.
Is it really there. In a
measurement of 171 technical domains,
6.3% returned HTTP 200 with an HTML body at /llms.txt
— the site catching an unknown path and rendering a page. Every one of those looks
fine to a status-code check and is useless to a model. This is the single most
common way an llms.txt is broken without anyone noticing.
Is it shaped like the spec. The proposal asks for an H1 with the site or project name, an optional blockquote summary, then H2 sections containing markdown link lists. A file with no links is not an index of anything.
Can anyone find it. There is no link relation and no
autodiscovery mechanism, so consumers must guess the well-known path. Only
7.5% of measured domains reference it from robots.txt,
which is currently the one place you can advertise it.
Is it served as text. A markdown file sent as
text/html invites the wrong handling downstream.
HTML at /llms.txt. Your framework's catch-all route is handling
the path. Serve the file as a static asset instead — put it in the directory your
host serves untouched (public/, static/,
dist/) rather than letting the router see it.
Not discoverable. Add a line to robots.txt. It is
not standardised, but it is where people look:
/ llms.txt: https://example.com/llms.txtNo links. An llms.txt without markdown links gives a model nothing to follow. List your real documentation URLs under H2 sections, with a short description after each link.