The Smart Dietary Assistant utilizes Machine Learning to provide personalized dietary advice, focusing on users with conditions like diabetes. This app leverages the Grounding DINO model, which combines a text encoder and image backbone to enhance food item detection without requiring a labeled dataset. With an AP score of 52.5 on the COCO dataset, the model demonstrates high accuracy in real-world scenarios, utilizing attention mechanisms to precisely recognize objects based on user-provided labels and images. Developed using React Native and TypeScript, the app operates seamlessly across multiple platforms and integrates a self-hosted PostgreSQL database, ensuring data integrity and enhancing user privacy. Key functionalities include personalized nutrition profiles, real-time food scanning, and health insights, facilitating informed dietary choices for health management and lifestyle optimization. Future developments aim to integrate wearable technologies for more tailored health recommendations. Keywords: Food Image Recognition, Machine Learning in Nutrition, Zero-Shot Object Detection