We are building an evidence-based AI Health Assistant designed to outperform generic chatbots by prioritizing accuracy and verification. Unlike standard AI that relies on general training data, our system uses Retrieval-Augmented Generation (RAG) to consult a trusted database of scientific papers and clinical guidelines before answering. It generates hyper-personalized, actionable nutrition plans that are strictly grounded in research. Key features include direct citations for every claim, redirects to source papers, and the ability to recommend specific study materials, ensuring users receive advice that is not only personalized but scientifically verified and hallucination-free.