The Grabette system provides a standardized framework for recording and sharing robot-manipulation data. By streamlining how researchers capture physical interactions, it reduces the friction of building large-scale imitation learning datasets. This open-source effort targets the data scarcity problem in robotics. Practitioners can now synchronize multi-modal sensor streams more reliably for model training.