Last October, I was sitting in a café in San Francisco watching a demo video on my phone. A robot arm was folding laundry. Not following a pre-programmed sequence—actually figuring out how to fold a shirt it had never seen before.
I must have watched it five times.
For eight years, we've been building motor controllers. Tinymovr started as a side project, grew into a real company, and our controllers now run in robots across 30+ countries—quadrupeds, arms, drones, exoskeletons. But watching that video, I realized something: the AI side of robotics is accelerating, and motion hardware needs to keep pace.
The researchers pushing embodied AI forward shouldn't have to become hardware experts just to run experiments. They need motion systems that work out of the box—reliable, affordable, well-documented.
That's why we're changing our name. Tinymovr becomes MotionLayer.
We're not stepping back from motion control. We're doubling down: EtherCAT controllers are in development, and the core product line keeps growing. But we're also building the tools that connect motion to learning: leader arms for teleoperation, open-source designs, hardware that integrates cleanly with the software stacks researchers already use.
We're starting with ML-101, an open-source leader arm with improved ergonomics, designed for VLA research. Fully open hardware and OSHWA certified, no licensing restrictions.
Build your own: https://github.com/tinymovr/ML101
If you're working on imitation learning, VLA models, or any flavor of embodied AI, reach out. We'd love to hear from you.
—Yannis
