TABASCAL (TrAjectory BAsed RFI Subtraction and CALibration) models and subtracts radio-frequency interference (RFI) in post-correlation radio-interferometry data by jointly fitting trajectories of RFI sources and antenna gain solutions. It builds a source-to-visibility forward model that incorporates near-field moving RFI (such as satellites or aircraft) and uses JAX (ascl:2111.002) and Dask to scale across CPUs and GPUs for calibration and subtraction tasks. The toolkit accepts visibility data, metadata on RFI trajectories or two-line elements, and YAML-based simulation configuration files to generate cleaned visibilities and produce diagnostics of gain and RFI fit quality. TABASCAL enables calibration workflows that recover astronomical signals with reduced data loss by explicitly modeling RFI trajectories rather than simply flagging contaminated data.