GList — AI-Powered Grocery App

Android grocery list app with AI receipt scanning, Gemini-powered item categorization, and spending analytics. Built with Kotlin and Jetpack Compose.

01.Project Overview

Overview

GList is a native Android grocery list application that combines on-device intelligence with cloud-based AI to transform how users manage groceries. The app uses the device camera to scan receipts, applies ML Kit for text recognition, and leverages Google's Gemini API to automatically categorize and organize grocery items.


Key Features

  • AI Receipt Scanning — Point the camera at a receipt using CameraX, ML Kit extracts text, and Gemini AI parses the items into structured grocery entries with categories, quantities, and prices
  • Smart Categorization — Items are automatically organized by category (produce, dairy, pantry, etc.) using Gemini's language understanding
  • Spending Analytics — Vico Charts visualize spending trends over time, category breakdowns, and budget tracking
  • Offline-First — Room database ensures the app works fully offline with local persistence
  • Material 3 Design — Modern, polished UI built entirely with Jetpack Compose and Material Design 3

Technical Architecture

The app follows a clean architecture pattern with dependency injection via Hilt:

  • UI Layer: Jetpack Compose with Material 3, Navigation Compose for routing
  • Data Layer: Room database for local persistence, DataStore for preferences
  • AI Layer: Google Gemini API for item categorization, ML Kit for OCR
  • Camera: CameraX for real-time receipt capture with Accompanist permissions handling
  • Monitoring: Firebase Crashlytics for crash reporting

Tech Stack

  • Kotlin, Jetpack Compose, Material Design 3
  • CameraX, ML Kit (Text Recognition)
  • Google Gemini AI
  • Room, Hilt, DataStore
  • Vico Charts, Firebase Crashlytics, Coil

Technologies

KotlinJetpack ComposeGemini AICameraXML KitRoomHiltFirebase

Role

Android Developer

Timeline

Jan 2026

Category

Mobile Development / AI