How it works
Paste your file and the checker walks line by line, flagging anything that violates the llms.txt draft specification. Errors are hard violations (missing H1, rule outside a User-agent block, path without a leading slash). Warnings are soft issues that most crawlers tolerate but some do not. Info messages are style nitpicks. Fix every error, skim the warnings.
What's checked
- Document structure. Exactly one # H1, blockquote directly underneath, heading levels nested correctly.
- Directive grammar. User-agent, Allow, Disallow, Sitemap, Last-Modified, Contact, License - each with its expected value format.
- Path syntax. Allow/Disallow paths must start with / or be empty; URLs in Sitemap must be absolute.
- Scope rules. Every Allow/Disallow must appear inside a User-agent block, not floating at the top of the file.
- File hygiene. No UTF-8 BOM, LF line endings preferred, no malformed markdown links.
Pair with
Syntax-clean? Run the full audit via the LLMs.txt Validator, preview how crawlers read the file with the LLMs.txt Preview, and check for robots.txt conflicts using the robots vs llms.txt checker. Strategy reading: what is llms.txt.