Despite the exponential growth of the amount of ground‐motion data, ground‐motion records are not always available for all distances, magnitudes, and site conditions cases. TFCGAN is a Python software package for modeling and simulating ground shaking to tackle this problem. Based on Esfahani et al. 2023, the software can be used as library in custom code or as command line application and can generate ground-shaking records in different domains (Fourier, Time-Frequency, and Time domains) and different formats (currently numpy, ascii, with foreseen implementation of other formats such as ASDF). The enclosed code and model consist of two steps. In the first step, the generative model simulates ground shaking by conditioning on a set of parameters. In the second step, the time-frequency domain is transferred to the time domain based on the phase retrieval algorithm. The model is conditioned on moment magnitude, distance, and shear wave velocity at the near-surface and trained using the KiK-net database. The proposed model is extended by using a hybrid dataset based on the combination of the European strong motion (ESM) database, near-fault ground-shaking records, and synthetic records. We validate our model based on terms of standard deviations for peak ground accelerations and Fourier amplitude spectral values.