We score your docs the way an agent sees them — can it find the page, parse the content, and use what it reads? Built on the AFDocs standard.
A live ranking of how every documentation site we track scores
Your docs have two readers now — the developer and the agent they sent ahead. Agents read at machine speed and skip past pages they can't parse.
Last month Agent requests tracked across docs
54,365,799_
Insights
Engineering and DevRel leaders who decided to stop guessing whether agents could read their docs.
Sarah Deaton
Technical Content Engineer, Anthropic
More and more people reach docs through agents. There's real upside in people engaging with content more dynamically with agents than a static page. However, if docs aren't agent ready, people get silent failures on surfaces nobody can see. Agent-ready docs are how we can ensure the upside of AI and your content is ultimately helping readers rather than misleading them.
Rhyannon Rodriguez
AI Researcher, Agent Ecosystem
Agents aren't web crawlers. They interact with docs differently than humans do, trimming content before it reaches their primary LLM. Pages built without solving for this constraint will consistently underperform. Your readers depend entirely on what an agent can extract from your public docs. Agent-friendly docs are how teams can close the gap and ensure the right information survives the pipeline intact.
Manny Silva
Maintainer, Doc Detective
AI multiplies documentation, for good and bad. One wrong code example no longer frustrates a handful of users — it delivers thousands of confidently wrong answers to every agent, RAG pipeline, and chatbot that asks. Your docs are now the source code of every AI product that reads them, and documentation debt has become a production-grade reliability problem.
