Lili-o Information

Lili-o builds a deterministic, on-edge brain for robots. Our first use case is shelf picking: a robot learns the task from a single demonstration and runs it reliably on a real warehouse or factory floor, with no months of data collection. It is built for deployability today, already running on Unitree and AgiBot humanoids. R&D in Paris, first market in Korea. Most robot pilots stall because they need long data collection and break when the real environment changes. Our approach learns from one demonstration and behaves predictably, so a task that works once keeps working. We have shown this on real automotive parts, and we ported the same brain to another humanoid platform in half a day. We are starting pilots with industrial partners and distributors in Korea and Europe. If you run picking or handling operations, or distribute robots and need them to work out of the box, we would like to talk.