We propose an AI-powered, wavelength-driven platform that accelerates the discovery of photosensitizers for photodynamic therapy (PDT). By training AI for porphyrin structure generation and automated spectral prediction, and a continuously expanding, traceable database, the system transforms a target wavelength into a ranked shortlist of high-quality molecular candidates with interpretable evidence. Researchers can rapidly identify clinically relevant photosensitizers matched to their light sources, reducing literature search and experimental screening from weeks to minutes. The platform combines predictive modeling, explainable molecular reasoning, and scalable web deployment to enable faster, more cost-effective PDT drug discovery.
