Prediction of realistic diffusion parameters of radionuclides in clay-rich host rocks and bentohite is relevant in the context of deep geological disposal of radioactive waste. The negative permanent charge of clay minerals (illite and smectite) enhances porewater concentration of aqueous cations and depletes that of anions near clay surfaces. This leads to greater diffusion rates of cations and to anion exclusion and retardation in compacted clayrock systems. The ClaySorDif model accounts for the effective diffusion coefficients De and distribution ratios Rd of elements and their aqueous species in argillaceous porous media. It extends the ClaySor model, which is a GEMS reincarnation of the widely known 2SPNE SC/CE multi-site sorption model by Bradbury and Baeyens (B&B), now made consistent with the PSI TDB 2020 chemical thermodynamic database. To help estimate De, Rd, and element-accessible porosity, the simple and efficient MPDL (mean potential in Donnan layer) model was implemented in Python within the ClaySorDif extension of pyGEMS/xGEMS codes. Verification of ClaySorDif against the PHREEQC results and the recent experimental data for Swiss argillaceous rocks shows that the simultaneous use of ClaySor and MPDL models is possible without any loss of accuracy. The modelling tool, running on the LES JupyterLab server, is currently in use for on-going research projects and applications.