What llms.txt is good for
llms.txt works as a guidance file for retrieval systems. It helps explain your product boundary, key public routes, and which resources should be cited before others.
For CitationGraph, llms.txt points crawlers toward the homepage, docs hub, blog, trust pages, and the limited GEO preview while explicitly warning against over-reading the public repository.
What llms.txt does not replace
It does not replace robots.txt, sitemap discovery, canonical tags, or structured data. Those remain the baseline for crawlability and page identity.
It also does not replace actual content architecture. If the site lacks stable docs and trust pages, llms.txt alone cannot create understanding.
How to structure it well
Use clear markdown headings, canonical public URLs, and explicit sections for preferred citation surfaces, trust pages, and product boundaries.
The file should be long enough to be useful, but specific enough that a retrieval system can understand what is public, what is invite-only, and what should not be inferred.
Key takeaways
- •llms.txt is guidance, not a shortcut.
- •Use it to map public routes and product boundaries clearly.
- •It works best when supported by strong docs, trust pages, and canonical metadata.