Alzheimer’s disease (AD) research is advancing rapidly, yet the growing complexity and volume of medical literature make it increasingly difficult for clinicians and researchers to access critical, evidence-based insights. We present ADDURE-Alzheimer’s Disease Document Understanding and Retrieval Engine, a domain-specific, multimodal Retrieval-Augmented Generation (RAG) framework designed to improve access to clinically relevant, evidence-based knowledge in AD. This system allows users to pose natural language queries and receive answers grounded in peer-reviewed literature, including content from text, tables, and figures. By leveraging optical character recognition (OCR), semantic enrichment, and a high-dimensional embedding architecture, ADDURE ensures that responses are both accurate and transparently linked to their source. The system was evaluated using a curated Alzheimer’s-focused benchmark (ADQA), where it achieved an overall accuracy of 89%, outperforming leading biomedical large language models and retrieval-based tools. ADDURE is actively being tested in clinical environments, where it supports a range of applications including decision support, automated literature review, and ongoing medical education. Its ability to bridge the gap between unstructured scientific knowledge and real-world clinical queries positions it as a valuable tool in the era of precision neurology and evidence-driven research synthesis.