llms.txt Adoption Is Mostly a Platform Default, Not a Decision
PlatformDTC Research
Abstract
The usual framing of llms.txt is that it is a cheap bet nobody has taken yet. Measured across 58 direct-to-consumer brands, 73% of reachable hosts already serve one — and 70% of those did not write it. The file most of them publish is not a summary document but a protocol handshake.
Key Findings
33 of 45 reachable brands (73%) serve a real llms.txt, rather than an HTML soft-404
23 of those 33 (70% of adopters) are a platform-generated template, clustering between 4.2KB and 4.4KB and opening with an identical line
10 of 33 are bespoke or use another format
13 of the 58 domains refused or failed to connect to an identified crawler
A Framing We Published and Then Disproved
The prevailing view — which this team published in the same week — is that no major assistant has publicly committed to consuming llms.txt, so publishing one is a small wager on a convention that might matter later. The data is not close to that. Nearly three quarters of reachable brands in the sample already serve one, which means the interesting question is not whether to adopt but who is actually authoring these files.
The File Sizes Gave It Away
Twenty-three of the thirty-three files cluster in a narrow band between 4.2KB and 4.4KB, and every one opens with the same generated header addressing AI agents and naming the brand's storefront. Sampling the underlying platform for those brands returned the same answer every time. The files are a platform default shipped to a merchant base, not thirty-three independent editorial decisions about how to talk to agents.
Why That Changes the Interpretation
An adoption percentage is normally read as a measure of intent — how many operators decided this mattered. When 70% of adopters did not author the file, the percentage measures platform market share instead. Any analysis that treats llms.txt adoption as a proxy for how seriously a category takes agent readiness is, on this sample, mostly measuring which storefront platform those brands happen to run.
Conclusion
Adoption statistics for a convention are only meaningful if the adopters chose it. Here most did not. The file is worth reading not as a declaration of intent but as evidence of what a platform has decided to say on its merchants' behalf — which is a different and, for a merchant, more urgent thing to know.
Published In Full
Read the complete study, method and datasetReferences
- [1]PlatformDTC Research (2026). Shopify Quietly Shipped Live Agentic Commerce to Its Whole Merchant Base. https://platformdtc.com/blog/llms-txt-adoption-study