GlucoTrack

GlucoTrack is a mobile-first app for people with Type 1 Diabetes. It helps users monitor glucose trends, log meals, estimate insulin units, and make smart choices at restaurants. The app uses a user-centered design and pairs accessible health features with a lightweight AI agent (“NutriTrackr”) and secure cloud services.

Features

  • Glucose monitoring
    Track blood sugar, visualize trends, and surface simple insights.

  • Smart food logging
    Snap a photo or enter food details. Get estimated carbs and insulin units with AI support.

  • Restaurant recommendations
    Search by ZIP code. See low-carb, lower-sugar options and quick swaps.

  • Predictive insights
    Spot recurring patterns, forecast likely highs or lows, and receive helpful reminders.

  • Accessibility first
    Friendly copy, manual input paths, offline-tolerant flows, and support for diverse cuisines.

  • Privacy and security
    Client-side protections, secure auth, encrypted storage where supported, and user control of data.

  • Cloud integration
    Cloud hosting, database, real-time APIs, and notifications for scale and reliability.


Tech Stack

Frontend: React, Tailwind CSS, shadcn/ui
Backend: Node.js + Express (HTTP API)
Database: Firebase Firestore (or Postgres via a managed provider)
Machine Learning: TensorFlow.js and/or hosted models
Cloud Services: Firebase, AWS, or GCP for hosting, auth, storage, and notifications
APIs: Nutrition and restaurant data sources for menu items and carb estimates
AI Agent: ai.py (Gemini-based assistant with USDA + web search tools)


Key files recap

  • ai/ai.py — Gemini-powered NutriTrackr agent. Pulls nutrition data from USDA FoodData Central, performs retrieval on past context, and answers in kid-friendly language.
  • ai/procedures.json — Curated procedures used to guide the agent’s step-by-step behavior.
  • frontend/src/components/RestaurantSearch.tsx — ZIP search and selection callback for restaurant suggestions.
  • backend/routes/aiChat.js — Bridges frontend chat input to the NutriTrackr agent.
  • backend/routes/restaurants.js — ZIP-based restaurant discovery and menu lookups.

Requirements

langchain-google-genai langchain langgraph chromadb python-dotenv requests

Tracks Applied

Product Design Cybersecurity Predict the Unpredictable Civics and Accessibility Healthcare Optimization

Awards Applied

Best Use of the Cloud Best Beginner

Team: BuckHacks

Krista Bair — kristalbair@gmail.com

Anika Talyan — anikatalyan@gmail.com

Joselyn Vasquez — joselynlilyanna@gmail.com

Melvin Vasquez — melvin.60834@gmail.com

Built With

Share this project:

Updates