NutriCare Agents

NutriCareAgents – Your intelligent food companion for personalized, healthy, and accessible meal planning

Comment

NutriCare Agents is an AI-powered, multi-agent nutritional advisory platform designed to deliver personalized, culturally relevant, and health-conscious food recommendations—beginning with Vietnamese users, and with long-term ambitions for global deployment. This project was developed for the GDGOC Hackathon Vietnam 2025 to tackle a significant public health challenge: how to make trustworthy, scientific, and personalized nutritional advice accessible to all, especially underserved and visually impaired populations.

The core innovation of NutriCare Agents lies in its multi-agent system architecture, where each agent serves a distinct role in a collaborative AI ecosystem:

  • A Meal Recommendation Agent uses Graph Neural Networks (GNNs) to model relationships between user health conditions, preferences, and local dishes—providing highly tailored suggestions.
  • A Knowledge Agent, powered by Google Gemini APIs and LangChain, explains the cultural significance, ingredients, and health value of each dish.
  • A Speech Agent, integrating OpenAI Whisper (STT) and Coqui TTS, enables voice-based interaction, allowing blind or elderly users to navigate the platform intuitively.
  • A Visual Agent processes images of food, using AI-powered dish recognition to identify meals and provide recommendations or cooking instructions.
  • A Location Agent, leveraging Google Maps and Places APIs, suggests nearby healthy food options.

We built NutriCare Agents using:

  • Frontend: React, Next.js, TailwindCSS.
  • Backend: Flask and Firebase Functions.
  • Database: MySQL for structured data, Firebase Realtime Database for user preferences.
  • Cloud: Azure for deployment and scalability.
  • AI Integration: LangChain, Gemini Pro, GNN, OpenAI Whisper, Coqui TTS.

Our team conducted surveys with over 400 users, which revealed critical insights: people desire healthy eating but lack access to culturally relevant, personalized advice. Many voiced a need for voice support and local food options that fit their budget and taste.

The impact of this system goes beyond nutrition: it enhances digital inclusivity, supports preventative healthcare, and respects cultural diversity in eating habits.

In the future, we plan to:

  • Expand our dataset to cover Southeast Asia and beyond.
  • Launch a mobile app with real-time health tracking.
  • Integrate data from wearable devices (e.g., smartwatches, glucose monitors).
  • Offer APIs for integration with healthcare providers and wellness apps.
  • Publish open-source modules to encourage community contribution and transparency.

NutriCare Agents is more than just a hackathon project. It’s a vision of Responsible AI for Real Lives—where personalization meets cultural understanding, and health meets accessibility.

Built With

Share this project:

Updates

Submission history