Robotic systems lack physics-aware training data for precise industrial assembly tasks involving force, torque, and tight-tolerance interactions, limiting real-world automation performance.

Juakali is a physics-based data pipeline that makes robot datasets with force feedback accessible to automate assembly planning.

It simulates fittings and fastening of mechanical parts from unstructured Computer-Aided Design (CAD) files while generating high quality structured datasets used to train Vision-Language-Action models that reduce time, cost and errors in engineering systems like Google Intrinsic Flowstate.

No expensive hardware like VR headsets are needed.

Using docker containers, enginers, developers and researchers or anyone with access to assembly dataset can install and run the containers in a matter of minutes.

These complex data can also be crowdsourced and users earn revenue.