Our robots operate in open scenarios.
Check out our robots operating in various real-world open scenarios, including supermarkets, warehouses and so on.
J-Box drives all robots.
J-Box is the cerebellum for any robot. It is universally applicable and can be plugged in to all forms of robots including wheeled, legged, and articulated ones. Our cross-embodiment behavior and motion control technology directly executes abstract behavior-level commands, like "go to somewhere", "pick up something", etc. It makes robot development as simple as plug-and-play.
J-Mind endows robots with intelligence.
J-Mind is a cloud-based large model platform, and supports both public and private deployment. It adds details to abstract goals from users and breaks down complex tasks in to fasible behavior sequences. Behind the scene is the incorporation of J-Layers (private scenario-/application-based contextual layers), LLM, LVM, L(X)M and multimodal foundation models.