Vision Lab is rebranding back to Linewise.
We originally started our journey as Linewise, building vision-based software for manufacturers, the "line" being the production line in a factory. Along the way we stumbled into a side project that turned out to matter more than the main one. The data that comes off a factory floor is exactly what robots and physical AI models need to learn from, and nobody was collecting or licensing it properly. So we built that. It grew faster than any of us expected, so we sunset the SaaS product and went all in.
We called that Vision Lab because it started with video data for robotics foundation model training. But we've outgrown the name. Our work now spans the full range of data that describes how physical things get made, not just what a camera sees. Anything you want to make, manufacture or fabricate leaves a data trail, and physical AI needs the whole value chain of producing a part. That's the path to Physical RSI: recursive self-improvement for the physical world, where robots and autonomous machines work together to design, build and improve things with minimal human hands on the line. We're a long way from it, but that's what we're building toward.
So we're back where we started, on the factory floor. Linewise is the right name for that. The line is where parts get made, and wise isn't only about knowledge. It's about knowing the right way to do it.
Machines are learning to build in the physical world. Everything they need to know is already there on the line, in the hands of the people who do it every day. We're just making sure none of it gets lost.
Linewise, the industrial data layer for Physical RSI.