The llms.txt guide for 2026.
llms.txt is the emerging convention for surfacing your most important content to AI assistants in a structured, model-readable way. Few sites have one. Early movers get the early signal. Here's what it is and how to deploy it correctly.
It's not robots.txt. It's the opposite.
Where robots.txt tells crawlers what they can't access, llms.txt tells AI assistants what's most important. Think of it as a curated index of your site optimized for LLM ingestion: clean markdown, hierarchical structure, links to canonical sources, no marketing fluff.
What a good llms.txt contains
A short site summary, key sections (Documentation, Product, About, etc.), and a curated list of your highest-value pages with one-sentence descriptions. Optional: a longer llms-full.txt with the actual content concatenated.
Which engines actually use it
Adoption is still partial. Some AI assistants check for llms.txt explicitly; others don't yet. Either way, deploying one is low-cost and signals to engines that you're optimized for AI consumption.
Reffed checks yours, Foundation will deploy it
Reffed checks whether your site has an llms.txt at the standard location and verifies it parses correctly. When Reffed Foundation ships in Q4 2026, it will generate llms.txt automatically based on your site structure and your most important pages, and update it as your content changes. Join the waitlist to lock in founding-member pricing.
It's not magic
llms.txt alone won't fix poor schema, bad content, or missing third-party citations. It's one signal in a much broader optimization stack. The businesses that benefit are the ones with strong foundations who add it as a polish.
Get your llms.txt deployed.
Free audit checks if your site has one and what it should contain.