Robotics is entering a new phase. Advances in AI and hardware are bringing machines out of controlled environments and into the messy, unstructured world, and new projects are appearing every week.
That progress is difficult to understand from the outside. Demos are made under different conditions, capabilities are described in different language, and the difference between a prototype, a pilot, and a deployed product is often unclear.
Machine Census gathers evidence about robots and the teams behind them: what each machine can do today, how it operates, and how far it has progressed. We organize that information into a common format and link claims to their sources, so every system can be understood in context.
Early work belongs in that picture. A prototype does not need to look commercial to be worth documenting. Teleoperation is a valid operating mode when it is disclosed. When information is missing, we mark it as unknown, not as failure.
The census is a living record. As machines improve, deployments grow, and new evidence appears, their profiles change with them.
We believe honest progress deserves to be visible, from the first working prototype to machines already at work in the world.