This paper presents an enhanced Retrieval-Augmented Generation (RAG) system designed specifically for orthopedic sports medicine. The system integrates multiple specialized knowledge indices, category-aware routing, and an agentic architecture to deliver precise and contextually appropriate information. By employing category-specific document processing and routing through specialized knowledge bases, the system provides medical practitioners with accurate responses tailored to specialized domains. We implement a multi-stage architecture including document categorization, specialized index creation, dynamic category routing, and response synthesis. Evaluation shows significant improvements in medical information accuracy and specificity compared to general-purpose RAG systems. The system represents a novel approach to domain-specific knowledge retrieval that connects foundation models with categorized domain knowledge, enabling more precise information access in specialized medical fields.