As robots become more present in open human environments, it will become\ncrucial for robotic systems to understand and predict human motion. Such\ncapabilities depend heavily on the quality and availability of motion capture\ndata. However, existing datasets of full-body motion rarely include 1) long\nsequences of manipulation tasks, 2) the 3D model of the workspace geometry, and\n3) eye-gaze, which are all important when a robot needs to predict the\nmovements of humans in close proximity. Hence, in this paper, we present a\nnovel dataset of full-body motion for everyday manipulation tasks, which\nincludes the above. The motion data was captured using a traditional motion\ncapture system based on reflective markers. We additionally captured eye-gaze\nusing a wearable pupil-tracking device. As we show in experiments, the dataset\ncan be used for the design and evaluation of full-body motion prediction\nalgorithms. Furthermore, our experiments show eye-gaze as a powerful predictor\nof human intent. The dataset includes 180 min of motion capture data with 1627\npick and place actions being performed. It is available at\nhttps://humans-to-robots-motion.github.io/mogaze and is planned to be extended\nto collaborative tasks with two humans in the near future.\n