What is llms.txt?

The llms.txt standard emerged in late 2025 as a practical mechanism for website owners to proactively communicate with AI search engines and large language model crawlers. The concept was pioneered by Answer.AI's Jeremy Howard and has been rapidly adopted by thousands of websites in the months since, particularly among technology companies and SaaS products that want to ensure accurate representation in AI-generated answers. Similar to how robots.txt tells crawlers what pages to access and not access, llms.txt tells AI systems what your site is about, which pages contain the most important information, and key facts that should be included in any AI-generated summaries or recommendations. The file is served at the root domain path (/llms.txt) in plain text format with a structured markdown-like syntax. A well-formatted llms.txt typically includes: a brief 2-4 sentence description of what the product or company does, section headers with bullet-point links to the most important pages in each category (product, blog, resources, company), and a 'Key Facts' section with specific, citable claims about the product (user counts, founding date, key differentiators). While llms.txt is not yet a formal web standard with browser or crawler enforcement, major AI platforms including ChatGPT's web search, Perplexity AI, and Anthropic's Claude have begun respecting the file in their crawl behavior. For SaaS companies, maintaining an accurate and comprehensive llms.txt ensures that AI search engines correctly represent your product's capabilities, pricing, and target audience when they generate answers about your category. Given that AI-powered search now influences discovery for a significant portion of B2B software purchases, this file has become an important part of any modern SEO and GEO strategy. The file should be treated as a living document and updated whenever significant product changes are made.