Open, MIT-licensed hypothesis-generation engine for MCAS/MCAD rescue, maintenance, and remission compounds. Pharma + herbs + supplements + AI-generated analogs, ranked transparently.

A reproducible, end-to-end computational pipeline that uses a machine-learning docking surrogate and receptor-aware fragment growing to design a new Nrf2-inhibitor candidate, followed by force-field parametrization, explicit-solvent energy minimization, and AI-assisted retrosynthetic assessment.