This project aims to embed intelligence into Building Information Modeling, benchmark with leading international standards, and establish an ecosystem based on a localized 3D graphics platform as the foundation application platform. Through self-develop innovative modules to provide independent and controllable information-based technology options for industry applications and educational research. We plan to propose automatic recognition conversion and generation techniques between 2D graphics and 3D models based on deep learning algorithms and automatic segmentation techniques for curtain walls. The functions developed in this project can be shared in the openBIM environment, and models or drawings generated through this project can be transferred with other software through IFC. The system will be a fundamental solution to the productivity problem, with its cloud-oriented BIM 3D graphics application environment features and the advantages of being highly adaptable to the openBIM environment. Simultaneously support the comprehensive application of BIM in the regional construction industry to achieve the productivity revolution.