Creating 3D scenes is a critical yet challenging digital content creation task due to the high demand for professional skills and intense labor efforts. Recently, significant progress has been made in accelerating 3D scene creation through AI assisted automated design and 3D content generation. However, current systems can not effectively support 3D creation workflow and inspire idea explorations. In this work, we present a novel 3D scene-creating copilot system that allows highquality and iterative 3D scene modeling and visualization. Under the hood, our system leverages a novel multi-step reasoning workflow to control language- and image-generative agents to create 3D scenes. Compared to traditional 3D modeling and visualization workflows, our system provides more natural and intuitive control and can significantly reduce the time required to create 3D content