AI Recipe Finder is a full-stack web application that suggests personalized recipes based on user-provided ingredients using NLP-powered prompts and Hugging Face LLMs. It integrates the Spoonacular API to fetch real-time recipe data, while cooking instructions are generated through the LLM for a more natural, conversational experience.
Users can filter results by diet, cuisine, or specific ingredients. Built with the MERN stack, the app combines efficient frontend/backend workflows with intelligent ML-driven responses. The core focus is on minimizing food waste and making meal planning easier using AI. The project demonstrates applied NLP, prompt engineering, and real-time ML system integration.
• Search recipes using the Spoonacular API. • Generate detailed cooking instructions with Hugging Face's Mistral-7BInstruct. • Backend built with Express, Node.js, and MongoDB. • Clean UI built with React.js and Axios. • Secure and sanitized backend with rate-limiting.
• React.js • Axios • Tailwind CSS (optional) • Vite (for fast dev server)
• Node.js • Express.js • MongoDB with Mongoose • Hugging Face Inference API (@huggingface/inference)
• Spoonacular API • Hugging Fac