This study presents a technology that can provide computer-based, signal-driven, musculoskeletal model-based analysis for musculoskeletal assessment post-stroke during gait. It comprises a modular 64 textile-embedded electromyography (EMG) leg garment integrated with a generalized and automated algorithm for the fast localization of leg muscles and an EMG-driven neuromusculoskeletal (NMS) modeling framework for the estimation of ankle torques and muscle forces. Our results showed that the automated clustering could extract muscle-specific clusters from 64 electrode activations of post-stroke individuals during a few gait cycles of a single walking task. Moreover, resulting muscle-specific EMG envelopes were electrophysiologically consistent for different walking speeds. Furthermore, the automated clustering provided EMG envelopes with the accuracy needed to drive an NMS modeling framework and estimate dynamically consistent ankle torques using unseen data. The technology proposed in this study opens new avenues for fast and quantitative musculoskeletal function assessment in neurologically impaired individuals.