What is llms.txt?

llms.txt is a Markdown-formatted file hosted at a website's root path that provides large language models with a structured, scannable summary of the site's key content, entities, and navigation. llms.txt was proposed as a standard by Jeremy Howard (co-founder of Answer.AI and fast.ai) to solve a specific retrieval problem: LLM context windows are too small to process entire websites, and converting complex HTML into machine-friendly text is slow and imprecise. llms.txt gives AI systems a curated entry point—listing entity names, descriptions, canonical URLs, and documentation links—so models can quickly understand what a site contains without crawling every page. In AI search optimization, llms.txt functions as a canonical source of truth that reduces hallucination risk by directing LLMs to verified entity data. llms.txt is one of three components of Entity API™ by Growth Marshal, alongside schema with graph properties and the brand fact file.