Team14 - Masala Mamu : Agentic AI Kitchen Assistant

Indian Institute of Science Summer 2025 DA225o Submission12 Authors

Keywords: Agentic AI, Kitchen, Cooking Assistant, Macro tracker, Recipe, Dishes, Nutrition, Health & Diet

TL;DR: A smart, AI-powered kitchen assistant that uses specialized agents and natural language to manage groceries, suggest recipes, track health, and optimize meal planning.

Abstract:

In the evolving landscape of smart homes, modern households face increasing complexity in managing day-to-day kitchen operations, including grocery tracking, meal planning, dietary health monitoring, and budget optimization. This project introduces the Agentic AI-Powered Kitchen Assistant, an intelligent, multi-agent system that automates and augments household kitchen management using the latest advancements in artificial intelligence, multimodal natural language processing, and external API integrations.

At the core of this system is a Router Agent (built using LangChain/LangGraph) that dynamically delegates tasks to specialized agents based on user queries and contextual requirements. These agents include:

Inventory Manager Agent: Maintains a real-time inventory by interfacing with a database and updates stock levels as items are consumed or added.
Price Comparison & Ordering Agent (MCP): Connects to multiple vendor APIs to identify cost-effective purchasing options and initiate grocery orders.
Recipe Generator Agent: Utilizes the current inventory and user preferences to propose recipes that are nutritious, and tailored to dietary restrictions.
Health & Diet Chart Agent: Tracks nutritional intake and provides personalized insights via a dashboard, backed by dietary guidelines.

The system communicates with a central LLM (Large Language Model) which facilitates natural language understanding and reasoning, bridging user intent with agent actions. The assistant supports multimodal NLP, allowing users to interact using both speech-to-text and text-based commands.

Submission Number: 12