Developing computer-assisted ligand design tools is key to accelerating drug discovery and studying molecular interactions. The article presents two complementary systems: LigandDesigner - HomeLab, optimized for laboratories with limited computational resources, and LigandDesigner - HPC, designed for high-performance supercomputing environments. Both systems use advanced machine learning models (LSTM, GAN) and integrate molecular dynamics, pharmacophore generation, and chemical validation techniques. The article presents the system architecture, predicted performance test results, and potential applications in life sciences.