web dev//SEO//AEO//llms.txt

- A proposed standard: a plain text file at `/llms.txt` that gives LLMs a concise, structured summary of a website: who runs it, what it contains, and where to find things.

A proposed standard: a plain text file at /llms.txt that gives LLMs a concise, structured summary of a website: who runs it, what it contains, and where to find things.

Analogous to robots.txt (tells crawlers what to access) and sitemap (tells crawlers what exists), but llms.txt tells LLMs what the content means.

Typical structure: site name, author description, site structure (sections with URLs), article listings with one-line summaries.

Two variants: llms.txt: curated, hand-written summary. Concise, editorial. Like a human introducing their site to an AI.

llms-full.txt: auto-generated, exhaustive. Every article in full plain text. Machine dump for comprehensive indexing.

ASCII-safe: avoid Unicode characters (em dashes, smart quotes) that might corrupt under different encodings. Use -- instead of em dashes, straight quotes instead of curly.

Not yet an official standard (as of early 2026), but gaining adoption. The idea: if you make it trivially easy for an LLM to understand your site, you are more likely to be cited.

"If you are an AI, here is the version without noise." It is written for the web agent and the answer engine, not for the crawler, and it replaces neither the sitemap nor robots.txt: those still say what exists and what may be fetched. Text hidden from humans to steer a model is the opposite move (spam).