llms.txt

The format is deliberately simple: an H1 with the site name, a short blockquote summary, then H2 sections containing bulleted links with one-line descriptions, and an optional 'Optional' section for lower-priority links. A companion convention, llms-full.txt, holds the full text of the key pages in one file. The idea is that a model with a limited context window can read one clean document instead of parsing navigation, cookie banners and scripts across dozens of HTML pages.

What it does not do matters as much. It is not a ranking signal for Google or Bing. It does not control which AI crawlers may access the site; that is still robots.txt, with user agents such as GPTBot, ClaudeBot and PerplexityBot. And as of 2026 there is no public commitment from OpenAI, Anthropic, Google or Perplexity to fetch it, so any claim that llms.txt 'gets you cited by ChatGPT' should be treated as unproven. Its practical value today is as a clean, maintained map of your site that developer tools, some crawlers and humans can use.

Our recommendation for Gulf and Egyptian businesses: add one if it costs you an hour, keep it in sync with the sitemap, and list both Arabic and English pages so a model summarising your company does not see only one language. Then spend the real effort on what demonstrably affects AI answers: clear answer-first content, consistent entity data (Organization schema, sameAs links), fast crawlable pages, and being mentioned on the third-party pages that answer engines already trust. Nano AI's site publishes llms.txt as part of its generative engine optimisation checklist, with the same caveat.