Check a site's llms.txt, or draft one from its pages.
Example
What an agent reads
Findings
llms.txt
Draft from the site
Capabilities and processing
Enter a website, or paste an llms.txt file. For a website, our server fetches its llms.txt, home page, robots.txt and sitemaps. A pasted file is read in your browser.
The file is checked against the llms.txt proposal from llmstxt.org: an H1 with the site's name first, an optional blockquote summary, optional notes, and H2 sections that list links as - [name](url): notes, with a section named Optional for links an agent may skip. The outline shows what an agent reads, each issue is marked on its line, and repairs make links absolute, put links in a list and turn extra H1 headings into sections. For a website, a draft llms.txt is made from what the site publishes: its name and description, the links in its navigation with their own text, and its sitemap, with pages known only from the sitemap named from their paths.
Example workflow
The sample shop's llms.txt writes its summary as plain text, uses one relative link, puts a link outside its list and has a list item without a link. The draft from its home page and sitemaps lists 18 pages in five sections. Break it removes the name, makes links relative and takes links out of their list, and shows a missing file and a file served as a web page.
Limits and interpretation
llms.txt is a proposal that some AI tools read; it does not affect search rankings or guarantee that an AI system uses it. Links are listed, not fetched. Review the draft's page names and add notes before publishing it. Nothing is stored.
Related tools
robots.txt and sitemap validator · Open Graph preview · Broken link checker · All tools · Privacy