What llms.txt is
A plain Markdown file at /llms.txt that gives AI crawlers a human-curated summary of your site: who you are, what you offer, where your most important pages live, and any disambiguating facts. It was proposed in 2025 and adopted broadly through 2026.
Why it exists
AI crawlers face a problem: your site may have hundreds of pages, some outdated, some promotional, some canonical. llms.txt lets you tell crawlers "here are the pages that matter and here is what they say". It is similar in spirit to robots.txt and sitemap.xml but written for AI consumption rather than classical search bots.
What a good llms.txt contains
A short H1 and tagline. A paragraph explaining who you are and who you serve. Linked lists of your core pages grouped by category (services, locations, pricing, guarantees). Contact info. Nothing promotional — just structured facts.
Which AI engines read it
As of late 2026, Perplexity reads it explicitly. ChatGPT's browsing agent uses it when present. Claude's agents respect it. Google's AI surfaces have not formally adopted it but experimental signals suggest they prefer sites that provide one.