Preparing AI Outputs for VR and AR

This chapter explores the translation of AI-generated imagery into immersive workflows for Virtual Reality (VR), Augmented Reality (AR), and Extended Reality (XR). While ComfyUI natively produces static images, these outputs can be extended into panoramic environments, depth-based parallax effects, and spatial overlays. The chapter presents a range of workflows, including the generation of equirectangular panoramas for VR navigation, depth map displacement for pseudo-3D illusions, AR projection mapping onto physical models, and hybrid methods that combine AI imagery with procedural geometry in Blender or Unity. Applications range from speculative mood environments to client presentations and design critiques, enabling users to “inhabit” atmospheres rather than view them as static renderings. Limitations such as perspective alignment, human scale calibration, and resolution mismatches are identified, while best practices for maintaining coherence across immersive platforms are provided. The discussion concludes with an outlook on AI-native XR, noting developments in NeRFs, Gaussian splatting, and multi-view pipelines that may soon converge with ComfyUI. AI imagery, when embedded into immersive media, becomes a dynamic scaffold for speculative design, transforming latent hallucinations into navigable experiential space.