A Context-Aware Medical Query System ( MediBot ): Bridging Large Language Models and Clinically Validated Documentation

The rapid advancement of Large Language Models (LLMs) has revolutionized the field of natural language processing, offering unprecedented capabilities in text generation and human-like interaction. However, in the high-stakes domain of healthcare, the deployment of generic LLMs is severely hindered by the “hallucination” phenomenon the tendency of models to generate factually incorrect or commercially biased information with high confidence [1]. This research introduces MediBot, a specialized medical chatbot architecture that utilizes Retrieval-Augmented Generation (RAG) to ground AI responses in clinically validated literature. The primary motivation behind MediBot is to address the data transparency gap. While models like GPT-4 or Llama 3 are trained on vast internet datasets, they often lack access to the most recent clinical guidelines or proprietary medical texts [2]. MediBot solves this by decoupling the knowledge base from the reasoning engine. By implementing a modular three-phase pipeline Knowledge Extraction, Vector Space Memory Setup, and Retrieval-QA Integration MediBot ensures that every answer provided to a user is derived from a specific, traceable source document.