MenuMatch is a mobile application that simplifies nutrition tracking for students at UConn dining halls using computer vision and modern AI models. Users log a meal by taking a photo of their plate, while the system captures context such as time and dining hall location to narrow down possible menu items. On the backend, MenuMatch queries our HuskyEats API to retrieve the relevant menu, and an LLM analyzes the image alongside these candidates to identify which foods are present. To estimate portion sizes, the system segments each food item and applies monocular depth estimation using MiDaS to approximate volume based on known plate geometry. An LLM then combines the identified foods, menu data, and volume estimates to infer serving sizes using density-based heuristics. Finally, the app retrieves the corresponding nutritional information and presents a detailed breakdown including calories, protein, fats, and carbohydrates, allowing students to quickly understand their meals without guesswork.