I built it twice. Didn't place twice. Then Google and Vercel shipped it.
A you.com rep said it at a hackathon kickoff in March. The web has a second audience now. The first one is people. The second is machines. When you ask ChatGPT or Perplexity something, the retrieval crawler behind it fetches your page and reads the markup that comes back. Not the site as a browser would draw it. The markup. That second audience idea really stuck with me. Then it fired me up. So I…
In March at a hackathon kickoff, a You.com representative emphasized that the web now has two audiences: people and machines. The idea of a second audience genuinely inspired the author to create a tool. Building it twice and submitting it to hackathons, neither version won, but on May 7, Google launched an Agentic Browsing category in Lighthouse, and on August 21, Vercel and Ora released is-agentic.com.
This led to four tools addressing the same question: can an agent use this site? The author's first attempt, Hermes, was a Python engine that scored well in six categories out of 100, but it missed two key aspects: reading files instead of pages and providing suggestions for improvement. The author rebuilt Agentis Lux, which runs on Perseus Clew and uses a live URL, focusing on six deterministic checks and raw HTML without executing JavaScript.
The author believes that findings are the foundation, not the entire structure, and that a fix is a decision made by someone who has never seen the code before. In May 2026, Google introduced the Agentic Browsing category in Lighthouse, which focuses on gathering data and providing actionable signals rather than a definitive ranking.
The category includes an llms.txt file, WebMCP, the agent-relevant slice of the accessibility tree, and cumulative layout shift. However, the scoring is not a weighted average, and the category is marked as experimental and under development. The author respects Google's restrained choice and understands that the measurement model is different from their own.
Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
