GaitPy is an open-source Python package that implements several published algorithms in a modular framework for extracting clinical features of gait from a single accelerometer device mounted on the lower back (L5 vertebra). The package has been developed to make it easy for researchers to derive measures of gait from raw accelerometer data. The package includes modules with three main functions: 1) classify bouts of gait; 2) extract clinical features of gait from each bout; and 3) visualize detected gait events.