# How we calculate the AI score Source: // Method The AI score measures how ready a site is to be read and cited by AI assistants and AI search. It runs from 0 to 100 and is the weighted mean of ten checks, computed for every working site we audited. | Check | Weight | A site passes when | | --------------------------------- | ------ | ---------------------------------------------------------------------------------------------------------------------------------------------------------------- | | Content in the HTMLpartial credit | 20 | its analyzed pages are not rendered by the browser alone. The score is the share of non client-side pages among those whose rendering is known. | | llms.txt | 15 | the file /llms.txt exists. | | robots.txt allows AI | 15 | robots.txt exists and does not disallow the whole site to every bot or to a known AI crawler. An Allow that re-opens part of the site is not taken into account. | | Structured data | 15 | the main page carries JSON-LD structured data. | | Sitemap | 10 | a sitemap.xml was found. | | Title | 5 | the main page has a title. | | Meta description | 5 | the main page has a meta description. | | Canonical | 5 | the main page declares a canonical URL. | | HTTPS | 5 | the site is served over HTTPS. | | Lighthouse SEOpartial credit | 5 | the score is the main page's Lighthouse SEO score, divided by 100. | | Total | 100 | | ## The AI crawlers we check for For the robots.txt check we look at whether the site is closed to every bot, or to any of these: - GPTBot - OAI-SearchBot - ChatGPT-User - ClaudeBot - Claude-Web - anthropic-ai - PerplexityBot - Google-Extended - CCBot - Applebot-Extended ## When a figure is missing A check the audit cannot answer is left out for that site, and the remaining weights are rescaled. A site whose page rendering is unknown, for instance, is scored out of 80 rather than given zero: missing data never counts as a failure. [Back to the observatory](https://www.nfw.run/state-of-the-web)