ensemble implements the Honeycutt (1992) inhomogeneous ensemble photometry method as an interactive Jupyter Lab widget to correct for atmospheric variations in differential light-curve photometry. It ingests background-subtracted photometry and uncertainties for a fixed set of stars across multiple exposures, solves for an ensemble solution, and lets users interactively mask stars, exposures, and individual star–exposure combinations via linked plots before recomputing the solution. The object-oriented design exposes attributes such as atmospheric-corrected instrumental magnitudes and masking matrices, enabling further analysis after the interactive session.