CalorieChat

AI-powered calorie tracking app with natural language meal logging.

Describe what you ate in plain English — the AI identifies foods, estimates portions, and calculates calories from USDA nutrition data.

---

Features

• Natural Language Meal Logging — Type "2 eggs and toast" and the AI parses it automatically
• Smart Food Matching — 100+ food synonyms and a 104-food USDA database
• Insights & Analytics — Daily summaries, 7-day trends, weekly averages
• Meal Editing — Update descriptions, adjust quantities, auto-recalculate calories
• Food Search — Wildcard patterns (*bread*, app*) with pagination
• CSV Export — Detailed meals or daily summaries, Excel/Google Sheets compatible
• Cross-Platform — Web (PWA), Windows desktop, Android/iOS (coming soon)
• Secure — API key stored server-side only, input validation, graceful error handling
• Offline Fallback — Local parser works when the backend is unavailable

---

Tech Stack

Layer	Technology
Frontend	Flutter 3.10+ / Dart, Provider, Freezed, Dio, sqflite
Backend	Node.js 18+, Express
AI	Google Gemini 2.0 Flash
Database	SQLite (mobile/desktop), in-memory (web)
Nutrition Data	USDA MyPyramid (104+ foods)

---

Quick Start

Prerequisites

• Flutter SDK 3.10.4+
• Node.js 18.0+
• Gemini API Key (free)

1. Clone the repo

git clone https://github.com/AaronBezi/CalorieChat.git
cd CalorieChat/calorie_chat

2. Start the backend

cd backend
npm install

Create backend/.env:

GEMINI_API_KEY=your_api_key_here
PORT=3000
NODE_ENV=development

npm start

3. Run the Flutter app (new terminal)

cd CalorieChat/calorie_chat

flutter pub get
flutter pub run build_runner build --delete-conflicting-outputs
flutter run -d chrome        # web
# or
flutter run -d windows       # desktop

Configuration

Flutter calorie_chat/.env:

API_BASE_URL=http://localhost:3000

Backend calorie_chat/backend/.env:

GEMINI_API_KEY=your_key_here
PORT=3000
NODE_ENV=development

---

Usage

Log a Meal

1. Open the Chat tab
2. Type a meal description: "2 eggs, toast, and coffee"
3. Click Parse Meal
4. Review detected foods and adjust quantities
5. Click Save Meal

Search Foods

1. Open the Search tab
2. Type a food name or use wildcards: *bread*
3. Browse paginated results

View Insights

1. Open the Insights tab
2. See today's totals and 7-day trends
3. Edit or delete meals
4. Export data to CSV

---

Architecture

calorie_chat/
├── lib/
│   ├── core/               # Config, networking, services, utils
│   ├── data/               # DAOs, models (Freezed), repositories
│   └── features/
│       ├── chat/           # AI meal logging
│       ├── search/         # Food search
│       └── insights/       # Analytics & history
├── backend/
│   └── server.js           # Express API proxy → Gemini
├── assets/data/            # USDA food database (JSON)
├── data_pipeline/          # USDA CSV-to-JSON conversion tools
└── test/                   # Unit tests (30 tests)

Data Flow

User Input → Flutter UI → Backend Proxy → Gemini AI
                 ↓                            ↓
          SQLite / Web Storage ← Parsed Results

---

Building for Production

# Web
flutter build web              # → build/web/

# Windows
flutter build windows          # → build/windows/runner/Release/

# Android
flutter build apk             # → build/app/outputs/flutter-apk/

---

Testing

flutter test

30 unit tests covering input validation, state management, food matching, search, pagination, and text normalization.

---

API

Health Check

GET /health
→ { "status": "ok", "timestamp": "...", "version": "1.0.0" }

Parse Meal

POST /api/parse-meal
Content-Type: application/json

{ "text": "2 eggs and toast" }

→ {
    "items": [
      { "query": "eggs", "quantity": 2, "portionHint": null },
      { "query": "toast", "quantity": 1, "portionHint": null }
    ]
  }

---