#What an llms.txt does
An llms.txt sits in the root of your website, just like robots.txt, and summarises in Markdown what a language model should know about your site. Where robots.txt controls what a bot may do, llms.txt says what matters. The structure comes from llmstxt.org: a title, a summary and sections with links. How to create the file, with an example, is covered in the article on llms.txt.
#What Flowproof checks
The scan fetches /llms.txt and first checks whether a real text file comes back. Many websites return an ordinary HTML page with status 200 for an unknown URL. To a bot the file then seems to exist, while it contains nothing usable. Next, Flowproof checks the structure line by line: an H1, a blockquote, H2 sections and the number of links. Finally, the scan looks at the Content-Type your server sends and whether there is an llms-full.txt.
Every finding comes with advice. If the file is missing, that needs attention and you get a template with the right structure. If it has too few links, you see the format to add them in.
#Honest about the impact
There is no public confirmation that ChatGPT, Gemini or Google AI Overviews use llms.txt when choosing sources. That's why Flowproof counts this area with a lower weight than AI crawler access or structured data. The file is quick to make and does no harm, but access for the bots and clear content on your pages achieve more. Also read about AI crawlers in robots.txt and the complete guide to GEO.
#Free in every scan
The llms.txt checker is one of the four core GEO areas you see in full with a free account, together with AI crawlers, technical GEO and schema markup. If you just want to test the file quickly, use the free llms.txt validator. For a complete assessment of your site, there's the GEO audit.