Skip to main content

Deployment

This guide covers deploying the AI Ingredient Scanner to production environments.

Architecture Overview​

Backend Deployment​

Google Cloud Run​

Cloud Run is recommended for the FastAPI backend.

1. Create Dockerfile​

FROM python:3.11-slim

WORKDIR /app

# Install dependencies
COPY requirements.txt .
RUN pip install --no-cache-dir -r requirements.txt

# Copy application
COPY . .

# Run with uvicorn
CMD ["uvicorn", "api:app", "--host", "0.0.0.0", "--port", "8080"]

2. Build and Push​

# Authenticate with GCP
gcloud auth configure-docker

# Build image
docker build -t gcr.io/YOUR_PROJECT/ingredient-scanner:latest .

# Push to Container Registry
docker push gcr.io/YOUR_PROJECT/ingredient-scanner:latest

3. Deploy to Cloud Run​

gcloud run deploy ingredient-scanner \
  --image gcr.io/YOUR_PROJECT/ingredient-scanner:latest \
  --platform managed \
  --region us-central1 \
  --allow-unauthenticated \
  --set-env-vars "GOOGLE_API_KEY=xxx,QDRANT_URL=xxx,QDRANT_API_KEY=xxx"

Environment Variables​

VariableRequiredDescription
GOOGLE_API_KEYYesGemini API key
QDRANT_URLYesQdrant Cloud URL
QDRANT_API_KEYYesQdrant API key
REDIS_URLNoRedis connection string
LANGCHAIN_API_KEYNoLangSmith API key
Security

Never commit API keys to version control. Use secret management:

  • Cloud Run: Secret Manager
  • Kubernetes: Secrets
  • Heroku: Config Vars

Alternative: Docker Compose​

For self-hosted deployments:

# docker-compose.yml
version: '3.8'

services:
  api:
    build: .
    ports:
      - "8000:8000"
    environment:
      - GOOGLE_API_KEY=${GOOGLE_API_KEY}
      - QDRANT_URL=${QDRANT_URL}
      - QDRANT_API_KEY=${QDRANT_API_KEY}
    restart: unless-stopped

  redis:
    image: redis:alpine
    ports:
      - "6379:6379"
    volumes:
      - redis_data:/data

volumes:
  redis_data:

Mobile App Deployment​

Expo EAS Build​

1. Install EAS CLI​

npm install -g eas-cli
eas login

2. Configure eas.json​

{
  "cli": {
    "version": ">= 3.0.0"
  },
  "build": {
    "development": {
      "developmentClient": true,
      "distribution": "internal"
    },
    "preview": {
      "distribution": "internal",
      "ios": {
        "simulator": true
      }
    },
    "production": {
      "ios": {
        "resourceClass": "m1-medium"
      },
      "android": {
        "buildType": "apk"
      }
    }
  },
  "submit": {
    "production": {}
  }
}

3. Update API URL​

Before building, update src/services/api.ts:

// Production API URL
const API_BASE_URL = 'https://your-cloud-run-url.run.app';

4. Build for Production​

# Android APK
eas build --platform android --profile production

# iOS (requires Apple Developer account)
eas build --platform ios --profile production

App Store Submission​

iOS App Store​

# Submit to App Store Connect
eas submit --platform ios --profile production

Requirements:

  • Apple Developer Program ($99/year)
  • App icons and screenshots
  • Privacy policy URL
  • App review information

Google Play Store​

# Submit to Google Play
eas submit --platform android --profile production

Requirements:

  • Google Play Developer account ($25 one-time)
  • App icons and screenshots
  • Privacy policy
  • Content rating questionnaire

Streamlit Deployment​

Streamlit Cloud​

  1. Push code to GitHub
  2. Connect at share.streamlit.io
  3. Configure secrets in Streamlit Cloud dashboard

Google Cloud Run​

# Dockerfile for Streamlit
FROM python:3.11-slim
WORKDIR /app
COPY requirements.txt .
RUN pip install --no-cache-dir -r requirements.txt
COPY . .
EXPOSE 8501
CMD ["streamlit", "run", "app.py", "--server.port=8501", "--server.address=0.0.0.0"]

Production Checklist​

Security​

  • API keys in secret management
  • CORS restricted to production domains
  • HTTPS enforced
  • Rate limiting configured
  • Input validation on all endpoints

Performance​

  • Qdrant Cloud in same region as API
  • Redis caching enabled
  • Connection pooling configured
  • Appropriate instance sizing

Monitoring​

  • LangSmith tracing enabled
  • Error alerting configured
  • Health check endpoints monitored
  • API latency tracked

Mobile​

  • Production API URL configured
  • App icons and splash screens
  • Privacy policy implemented
  • Crash reporting (Sentry/Crashlytics)

Scaling Considerations​

API Scaling​

Cloud Run auto-scales based on requests. Configure:

  • Min instances: 1 (avoid cold starts)
  • Max instances: Based on budget
  • Concurrency: 80 requests per instance

Qdrant Scaling​

For high-volume usage:

  • Use Qdrant Cloud's distributed mode
  • Configure read replicas
  • Consider dedicated cluster

Cost Optimization​

ServiceFree TierTypical Cost
Google Gemini15 RPMPay-per-use
Qdrant Cloud1GB free$25+/month
Cloud Run2M requestsPay-per-use
Redis Cloud30MB free$5+/month