llms-full.txt
llms-full.txt is a community convention for publishing an expanded, single-file version of a site's content for language models to consume, alongside the shorter index-style llms.txt. It is not part of the original llms.txt proposal, which describes optional expanded files generated by tooling rather than a standard filename, and like llms.txt it is honoured by no major AI provider as a matter of policy.
What the proposal actually says
Jeremy Howard proposed llms.txt in September 2024: a markdown file at the domain root giving models a curated map of a site's most useful content.
The original proposal does not define a filename called llms-full.txt. It describes generating expanded context files from an llms.txt using tooling, and the example given uses names like llms-ctx.txt and llms-ctx-full.txt. The llms-full.txt name grew out of adoption by documentation platforms and became a de facto convention rather than a specification.
Worth knowing before someone cites it to you as a standard. It is a widely-copied habit.
Does anything read it?
The evidence available is not encouraging. Ahrefs analyzed 137,210 domains with traffic in May 2026 and found that 28% published an llms.txt file while 97% of those files received no traffic at all that month, with retrieval bots such as PerplexityBot and OAI-SearchBot accounting for around 1.1% of the requests that did arrive.
No major AI provider has committed to reading either file. Neither grants nor blocks anything — that is robots.txt's job.
The format has found a real niche in developer documentation, where a full-text file is genuinely convenient for a developer to paste into a model's context window. That use case is human-initiated and works. It is not the same as an AI search engine discovering and preferring your content.
Should a catalog publish one?
Almost certainly not in the full-content form, and for a reason specific to catalogs rather than a general objection.
A distributor catalog is hundreds of thousands of product records. A single file containing them would be enormous, immediately stale on price and stock, and duplicative of the feeds and structured data that already carry that information in formats consuming systems actually use. There is no mechanism by which it improves retrieval.
If you want to publish something, a short llms.txt pointing at your key category and guide pages costs an afternoon and does no harm. It just is not the reason your SKUs are not being cited. Fill rate and rendering are.
Frequently asked questions
What is llms-full.txt?
A community convention for a single expanded file containing a site's full content for language models, published alongside the shorter llms.txt. It is not defined in the original llms.txt proposal, which instead describes generating expanded context files with tooling under different names.
Do AI companies read llms.txt or llms-full.txt?
No major provider has committed to it. An Ahrefs study of 137,210 domains with traffic in May 2026 found 28 percent published an llms.txt file while 97 percent of those files received no traffic that month, with retrieval bots making up around one percent of requests that did arrive.
Should an ecommerce site publish llms-full.txt?
For a large catalog, no. The file would be enormous, stale on price and availability the moment it was written, and duplicative of feeds and structured data that consuming systems actually use. A short llms.txt pointing at key pages is harmless if you want one.
Is llms.txt a replacement for robots.txt?
No. robots.txt controls crawler access and is honoured by the documented agents from major providers. llms.txt and llms-full.txt grant and block nothing; they are content hints with no enforcement and no committed consumer.