DescriptionWith the prevalence of neural networks and deep learning models, more data is required to expand the domain as well as to improve the accuracy of those models. There are numerous annotation tools and software of RGB data aiming to make the labeling process less gradual and more efficient while maintaining the same accuracy as traditional methods. However, fewer such efforts have been made in the RGB-D domain. This paper provides a novel RGB-D annotation tool that is designed to efficiently generate object poses in images or video sequences. The tool is equipped with functions, such as removing background points, interactive marker, to aid annotation, as well as ICP to lower the number of frames that need to be labeled in a video sequence.