In this work we demonstrate an AI-based exoskeleton controller that is capable of accurately capturing both internal and external states of the user. This allows for assistance modulation that is closely matched to the user’s varying biomechanical demand during walking in real-world without any user specific tuning. The assistance using our intelligent controller was more preferred to navigate the community with and more effective in offloading the user’s energetic consumption compared to the conventional method. The training data used for the models and the CAD design of the exoskeleton hardware used in this study are open-sourced alongside the publication for researchers to leverage these resources for their exoskeleton developments. This work is an exciting stride toward the deployment of an intelligent exoskeleton for assisting human locomotion in real-world community ambulation.