New open-source standard defines technical baseline for web development
Platform-agnostic document outlines ten key areas for site construction, including accessibility, security, and AI agent legibility, with sources credited to major standards bodies.
A platform-agnostic technical standard for web development, titled The Website Specification, has been published under an MIT licence. The document outlines ten key areas for website construction, including search visibility, accessibility aligned with WCAG, security, AI agent legibility, performance, consent, and internationalisation. The project is developed openly on GitHub, with contributions managed via pull requests.
The specification is designed to be framework-agnostic, applying equally to static site generators such as Astro and Hugo, web frameworks like Next.js and Django, and content management systems including WordPress and Drupal. Implementation hints are provided separately from the core specification to ensure that the standards remain primary. Each topic within the document links back to its source standard, crediting organisations such as the Web Hypertext Application Technology Working Group (WHATWG), the World Wide Web Consortium (W3C), and the Internet Engineering Task Force (IETF).
The project is accessible as a read-only, open Model Context Protocol (MCP) server that requires no authentication. It also includes a published Agent Skill designed to teach compatible AI agents when and how to utilise the specification. Additionally, the document is available in Markdown format via `/llms.txt` and the `Accept: text/markdown` header on any spec URL.
A checklist is provided for evaluating whether a site adheres to the standard, with each item offering implementation guidance. The checklist allows users to run through ten items, determining if a site performs specific functions with a simple yes or no response. Clicking into any item reveals what the standard is, why it matters, and how to implement it.
The specification covers technical requirements ranging from HTML basics and search visibility through to headers, transport policies, and graceful failure handling for error pages and offline states. It also addresses language, locale, and direction for translated content. The project credits specific standards bodies for each topic, including WHATWG, W3C, IETF RFCs, WCAG, and MDN, ensuring that the document reflects the organisations defining the modern web.
