Accurate regional and seasonal prediction of marine ecosystems requires robust and flexible modelling tools. We introduce py-off-bgc, a Python-based modelling framework designed to simulate ocean biogeochemistry offline. By utilizing prescribed daily-mean 3D physical fields (temperature, salinity, and velocity) from established ocean reanalysis products and forecasts (e.g., JCOPE, LORA, BRAN, CMEMS, SINTEX-F), py-off-bgc ensures realistic tracer transport without the computational overhead of coupled physics. This offline architecture allows us to focus on refining complex ecological processes and developing targeted data assimilation methods. In this presentation, we showcase preliminary results from regional py-off-bgc simulations and discuss the framework's trajectory toward seasonal prediction.