About Docs for Agents
Docs for Agents publishes the Docs Agent-Readability Test: a recurring reading test that asks whether a product’s public documentation can get an AI agent through five first-hour developer jobs. Each run produces a scorecard, a published report, and at least one verified receipt.
The premise is that documentation now has two audiences. Developers still read the docs, and increasingly their AI agents read the docs first: fetching quickstarts, reconciling rate-limit pages, and writing integration code from whatever the pages actually say. The test measures how well that second audience is served.
Reports live on this site and the series runs on LinkedIn. Every run follows the same battery, panel, and verification rules, documented on the methodology page.
Who runs it
Sean Knight is a serial entrepreneur in the San Francisco Bay Area who builds products and companies around AI and geospatial data.
He started out as an astrophysicist, pivoting to remote sensing, geospatial data science, and ML. Eventually that had him flying drones over glaciers in Alaska and New Zealand to build 3D point clouds.
He also started a retail computer shop at age 17, interned at NASA JPL, and has done research at a synchrotron and a research nuclear reactor.
These days his focus is on starting companies, AI consulting, and building AI agents that run geospatial pipelines on platforms like Databricks.