The increasing demand for accessible and continuous healthcare has motivated the development of intelligent digital health platforms. This paper presents MedConnect, an AI-powered health companion that integrates real-time vitals monitoring using IoT devices, simplification of complex medical reports using large language models, interactive 3D anatomy visualization, and an automated emergency response system. Users can upload medical reports in PDF, image, or text format. The system employs Optical Character Recognition (OCR) with Tesseract, Medical Named Entity Recognition using SpaCy and Med7, and a locally hosted LLaMA-3 8B model via Ollama to generate easy-to-understand plain-language summaries. Identified conditions are highlighted on an interactive 3D human anatomy model built with Three.js and React-Three-Fiber. The platform further supports voice input and output in English, Telugu, and Hindi, real-time emotion detection through webcam, and a reliable SOS mechanism that auto-dials 108 while alerting registered caretakers. Evaluation on a custom dataset of 150 de-identified medical reports yielded a SARI score of 47.6, BERTScore of 73.2, anatomy-mapping accuracy of 91.4%, and emergency response reliability between 96–100%. MedConnect offers a low-cost, privacy-preserving, and scalable solution suitable for preventive healthcare in resource-constrained settings such as India.