This study proposes a new method for creating and sharing remote collaboration spaces using neural radiance fields (NeRF) in extended reality (XR) environments. This method aims to enable remote users to collaborate each other in XR-based environments that closely resemble on-site conditions, thus facilitating intuitive interactions and exploring the potential to save time and costs across various industries. The proposed approach trains the NeRF model using RGB images and camera pose information collected from on-site workers' devices to generate 3D virtual and volumeric spaces, which are then visualized in real-time in the virtual environments of remote users. In particular, the NeRF-based model is free from optical limitations, such as light reflec- tion, transmission, and sensor range limitations. Thus, the proposed approach can be versatile and applicable in diverse environments without the need for depth sensors. In addition, it allows for fast spatial reconstruction and sharing via smart glasses and mobile devices, making it adaptable not only to fixed working environments like simple machine inspection but also to dynamic working situations, such as construction and manufacturing tasks. The proposed system demon- strates the potential to effectively create NeRF-based...