KaloriN AI

KaloriN AI is a nutrition tracking web app with AI-powered food scanning, meal logging, personalized recommendations, BMI-based calorie targets, and weekly nutrition insights.

.
├── frontend/              # React app
├── backend/               # Express API and Prisma schema
├── ai_service/            # FastAPI AI microservice and ML assets
├── APP_LOGIC_DOCUMENTATION.md
├── BACKEND_DOCUMENTATION.md
└── USER_GUIDE.md

Install these before running the full app:

• Node.js 18+ and npm
• Python 3.10+
• Supabase PostgreSQL project
• Firebase project with Authentication enabled
• Gemini API key
• Redis (optional, recommended for caching)

Create local .env files in each service. Do not commit real secrets.

VITE_API_URL=http://localhost:5000

Firebase config is currently set in frontend/src/config/firebase.js.

PORT=5000
DATABASE_URL=postgresql://USER:PASSWORD@HOST:PORT/postgres
DIRECT_URL=postgresql://USER:PASSWORD@HOST:PORT/postgres
AI_URL=http://localhost:8000
AI_TIMEOUT=15000
REDIS_URL=redis://localhost:6379
FIREBASE_SERVICE_ACCOUNT_BASE64=base64_encoded_firebase_service_account_json
CORS_ORIGINS=http://localhost:5173,http://127.0.0.1:5173
SUPABASE_URL=https://PROJECT_ID.supabase.co
SUPABASE_SERVICE_ROLE_KEY=your_supabase_service_role_key
SUPABASE_AVATAR_BUCKET=avatars

# Optional external recipe APIs
SPOONACULAR_API_KEY=
EDAMAM_APP_ID=
EDAMAM_APP_KEY=

Notes:

• FIREBASE_SERVICE_ACCOUNT_BASE64 is used by backend to verify Firebase ID tokens.
• CORS_ORIGINS is a comma-separated allowlist.
• SUPABASE_SERVICE_ROLE_KEY must remain server-side only.

GOOGLE_API_KEY=your_gemini_api_key
REDIS_URL=redis://localhost:6379
GEMINI_TIMEOUT_SECONDS=12

git clone https://github.com/ryhndastra/kalorin-ai.git
cd kalorinAi

cd backend
npm install
npx prisma generate
npx prisma migrate dev
npx prisma db seed
npm run dev

Backend runs at http://localhost:5000.

cd ai_service
python -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
uvicorn main:app --reload --port 8000

AI service runs at http://localhost:8000.

cd frontend
npm install
npm run dev

Frontend runs at http://localhost:5173.

redis-server

If Redis is not running, the app still works but AI caching is reduced.

Protected routes are guarded in the frontend.

More backend details: BACKEND_DOCUMENTATION.md.

Prisma schema: backend/prisma/schema.prisma

Main models:

• Profile
• Food
• DailyLog
• DailyInsight

Seed data: backend/data/nutrition_data.json

Run services in separate terminals:

# Terminal 1
cd backend
npm run dev

# Terminal 2
cd ai_service
source .venv/bin/activate
uvicorn main:app --reload --port 8000

# Terminal 3
cd frontend
npm run dev

• Set up modular architecture (frontend, backend, ai_service)
• Implemented Firebase authentication (email/password + Google)
• Added profile onboarding and BMI/calorie/protein target calculation
• Built AI food scanner integration
• Built food recommendation and explanation flow
• Added daily tracking and weekly insight modules
• Added guest mode for limited access

• No end-to-end test suite yet
• Some external API integrations are optional and key-dependent
• Redis is optional but affects recommendation/explanation caching quality

• APP_LOGIC_DOCUMENTATION.md - app flow and architecture notes
• BACKEND_DOCUMENTATION.md - backend implementation details
• USER_GUIDE.md - non-technical usage guide

# Frontend
cd frontend
npm run dev
npm run build
npm run lint

# Backend
cd backend
npm run dev
npx prisma studio
npx prisma db seed

# AI service
cd ai_service
uvicorn main:app --reload --port 8000

• Authenticated users must complete birthdate, weight, and height before using protected features.
• Guest users can access only landing and food analysis pages.