Deployment Cheat Sheet

Your complete reference for deploying AI apps. Bookmark this page for quick access to commands, templates, and checklists.

Platform Comparison Table

Quick Recommendations

• Next.js + AI → Vercel
• React + Firebase backend → Firebase Hosting
• Static site → Netlify or Vercel
• Python/Express API → Railway or Render
• Full control needed → Railway

Deployment Command Reference

Vercel

# Install
npm install -g vercel

# Login
vercel login

# Deploy (preview)
vercel

# Deploy (production)
vercel --prod

# Set environment variable
vercel env add VARIABLE_NAME

# List deployments
vercel ls

# Remove a deployment
vercel rm deployment-url

Firebase

# Install
npm install -g firebase-tools

# Login
firebase login

# Initialize
firebase init

# Deploy everything
firebase deploy

# Deploy only hosting
firebase deploy --only hosting

# Deploy only functions
firebase deploy --only functions

# Deploy only Firestore rules
firebase deploy --only firestore:rules

# Preview channel
firebase hosting:channel:deploy preview-name

# Set function secret
firebase functions:secrets:set SECRET_NAME

# Start emulators
firebase emulators:start

# Switch project
firebase use project-id

Netlify

# Install
npm install -g netlify-cli

# Login
netlify login

# Deploy (preview)
netlify deploy

# Deploy (production)
netlify deploy --prod

# Set environment variable
netlify env:set VARIABLE_NAME value

# Open site
netlify open

Railway

# Install
npm install -g @railway/cli

# Login
railway login

# Link to project
railway link

# Deploy
railway up

# Set variable
railway variables set KEY=value

# Open dashboard
railway open

# View logs
railway logs

Environment Variable Checklist

Required for Most AI Apps

# AI API Keys (server-side only — never prefix with NEXT_PUBLIC_ or VITE_)
OPENAI_API_KEY=sk-...
ANTHROPIC_API_KEY=sk-ant-...
GOOGLE_AI_API_KEY=...

# Firebase (client-side — use public prefix)
NEXT_PUBLIC_FIREBASE_API_KEY=...
NEXT_PUBLIC_FIREBASE_AUTH_DOMAIN=...
NEXT_PUBLIC_FIREBASE_PROJECT_ID=...
NEXT_PUBLIC_FIREBASE_STORAGE_BUCKET=...
NEXT_PUBLIC_FIREBASE_MESSAGING_SENDER_ID=...
NEXT_PUBLIC_FIREBASE_APP_ID=...

# App Configuration
NEXT_PUBLIC_APP_URL=https://your-app.com
NODE_ENV=production

# Error Tracking
NEXT_PUBLIC_SENTRY_DSN=...
SENTRY_AUTH_TOKEN=...

# Database (if not using Firebase)
DATABASE_URL=...

.env.example Template

# AI APIs
OPENAI_API_KEY=your-openai-key
ANTHROPIC_API_KEY=your-anthropic-key

# Firebase
NEXT_PUBLIC_FIREBASE_API_KEY=your-firebase-api-key
NEXT_PUBLIC_FIREBASE_AUTH_DOMAIN=your-app.firebaseapp.com
NEXT_PUBLIC_FIREBASE_PROJECT_ID=your-project-id
NEXT_PUBLIC_FIREBASE_STORAGE_BUCKET=your-app.firebasestorage.app
NEXT_PUBLIC_FIREBASE_MESSAGING_SENDER_ID=your-sender-id
NEXT_PUBLIC_FIREBASE_APP_ID=your-app-id

# App
NEXT_PUBLIC_APP_URL=http://localhost:3000

# Monitoring (optional)
NEXT_PUBLIC_SENTRY_DSN=your-sentry-dsn

CI/CD Template

Complete GitHub Actions Workflow

# .github/workflows/ci-cd.yml
name: CI/CD

on:
  push:
    branches: [main]
  pull_request:
    branches: [main]

jobs:
  test:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
      - uses: actions/setup-node@v4
        with:
          node-version: "20"
          cache: "npm"
      - run: npm ci
      - run: npm run lint
      - run: npx tsc --noEmit
      - run: npm test
      - run: npm run build
        env:
          NEXT_PUBLIC_APP_URL: https://example.com

  deploy:
    needs: test
    if: github.event_name == 'push' && github.ref == 'refs/heads/main'
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
      - uses: actions/setup-node@v4
        with:
          node-version: "20"
          cache: "npm"
      - run: npm ci
      - run: npm run build
        env:
          NEXT_PUBLIC_APP_URL: https://your-app.com
      # For Firebase:
      - uses: FirebaseExtended/action-hosting-deploy@v0
        with:
          repoToken: ${{ secrets.GITHUB_TOKEN }}
          firebaseServiceAccount: ${{ secrets.FIREBASE_SERVICE_ACCOUNT }}
          channelId: live
          projectId: your-project-id
      # For Vercel (alternative):
      # - uses: amondnet/vercel-action@v25
      #   with:
      #     vercel-token: ${{ secrets.VERCEL_TOKEN }}
      #     vercel-org-id: ${{ secrets.VERCEL_ORG_ID }}
      #     vercel-project-id: ${{ secrets.VERCEL_PROJECT_ID }}
      #     vercel-args: "--prod"

Cost Optimization Checklist

Before Launch

✅ Set spending limits on all AI API accounts
✅ Choose the cheapest model that meets quality needs
✅ Implement response caching (in-memory or database)
✅ Set max_tokens on all AI API calls
✅ Add rate limiting per user
✅ Test with production-like load to estimate costs

Ongoing

✅ Review token usage weekly
✅ Monitor cost per user
✅ Check cache hit rate (target: 20%+)
✅ Watch for cost spikes in daily reports
✅ Evaluate newer, cheaper models as they release
✅ Trim prompt templates to reduce input tokens

Cost-Saving Techniques

Pre-Deployment Checklist

Code Quality

✅ All tests passing
✅ No TypeScript errors
✅ No linting warnings
✅ Build succeeds without errors
✅ No console.log statements in production code
✅ Error boundaries in place for React components

Security

✅ All secrets in environment variables (not in code)
✅ .env in .gitignore
✅ AI API keys are server-side only
✅ Firestore security rules are NOT in test mode
✅ Storage security rules configured
✅ CORS configured correctly
✅ Rate limiting active
✅ Input validation on all user inputs

Performance

✅ Images optimized (WebP, proper sizes)
✅ Caching headers set for static assets
✅ AI response caching implemented
✅ Loading states for AI operations
✅ Error handling for slow/failed AI calls
✅ Streaming enabled for long AI responses

Monitoring

✅ Error tracking set up (Sentry or similar)
✅ AI API call logging active
✅ Cost tracking in place
✅ Health check endpoint available
✅ Alerts configured for error spikes and cost overruns

AI Prompts for Deployment Help

Initial Setup

• "Help me deploy my [framework] app to [platform]. Walk me through each step."
• "Configure my project for production deployment. Here's my tech stack: [list tech]."
• "Set up environment variables for my AI app on [platform]. Here are the variables I need: [list them]."

CI/CD

• "Create a GitHub Actions workflow that tests and deploys my [framework] app to [platform]."
• "My CI pipeline is failing with this error: [paste error]. How do I fix it?"
• "Add preview deployments to my workflow so every PR gets a preview URL."

Monitoring

• "Set up error tracking with Sentry for my [framework] app."
• "Create a logging wrapper for my AI API calls that tracks latency, tokens, and costs."
• "Build a simple monitoring dashboard for my AI app's usage and costs."

Cost Optimization

• "My AI app costs $[X]/month. Here's how it makes API calls: [describe]. How can I reduce costs?"
• "Implement caching for my AI API calls to reduce costs. Use [Firestore/Redis/in-memory]."
• "Help me choose the right AI model for each feature in my app: [list features]."

Troubleshooting

• "My app works locally but not in production. Here's what happens: [describe]. Build logs: [paste]."
• "Users are reporting slow AI responses. How do I diagnose and fix this?"
• "My deployment keeps failing. Here's the error: [paste]. Here's my config: [paste]."
• "I accidentally committed my API key to Git. What should I do?"

The Complete Deployment Workflow

1. DEVELOP locally with .env file
2. TEST with npm run lint, tsc, npm test
3. BUILD with npm run build
4. CONFIGURE environment variables on your hosting platform
5. DEPLOY to a preview environment first
6. VERIFY the preview — test AI features, check error tracking
7. DEPLOY to production
8. MONITOR logs, errors, costs, and performance
9. ITERATE — review monitoring data, optimize costs, improve reliability

Quick Reference: Common Errors and Fixes

---

Congratulations! You now know how to deploy, monitor, and scale AI-powered applications. The key takeaways:

1. Start simple — Deploy to Vercel or Firebase with one command
2. Automate early — Set up CI/CD before your first user
3. Monitor everything — Especially AI API costs and latency
4. Cache aggressively — It's the easiest way to cut costs
5. Scale gradually — Start with cheaper models and upgrade as needed

Happy deploying!

---

🌐 www.genai-mentor.ai