Skip to content
Kernova

Step 6: Deploy to Production

Step 6: Deploy to Production

Your support bot works locally. Let’s put it on the internet.

Sign Up for Kernova Cloud

  1. Go to kernova-cloud.vercel.app
  2. Create an account
  3. Go to Dashboard → Settings
  4. Click “Create Key” to generate an API key
  5. Copy the key (it starts with kn_live_)

Login from CLI

Terminal window
kernova login --key kn_live_your_key_here

You’ll see:

  ✓ Logged in to Kernova Cloud
    Key: kn_live_xxxxxxxx...
    Cloud: https://kernova-cloud.vercel.app

Add Your Secrets

Before deploying, add your OpenAI key to the cloud vault so your production deployment can access it:

  1. Go to Dashboard → Secrets
  2. Click “Add Secret”
  3. Key: OPENAI_API_KEY, Value: your key
  4. Save

This key will be injected into your deployment as an environment variable.

Deploy

Terminal window
kernova deploy

The CLI will:

  ┌──────────────────────────────────────────┐
  │  🚀 Kernova Full-Stack Dev Server        │
  └──────────────────────────────────────────┘


  Deploying: frontend + agents
  Region:    us-east-1


  [1/4] Validating configuration...  ✓
  [2/4] Compiling agents...          ✓
  [3/4] Building frontend (Next.js)... ✓
  [4/4] Uploading to Kernova Cloud...


  ┌──────────────────────────────────────────────────┐
  │  ✅ Deployed successfully                        │
  │                                                  │
  │  URL:     https://support-bot.kernova.app        │
  │  Region:  us-east-1                              │
  │  Status:  live                                   │
  └──────────────────────────────────────────────────┘

Verify It’s Working

Open your deployed URL:

Terminal window
# Check the health endpoint
curl https://support-bot.kernova.app/v1/health
{
  "status": "healthy",
  "name": "support-bot",
  "agents": [{ "route": "support", "name": "Support Agent" }],
  "uptime": 42.5
}

Visit https://support-bot.kernova.app/chat — your chat UI is live.

Test via API

You can also call the agent directly via API:

Terminal window
curl -X POST https://support-bot.kernova.app/api/agents/support \
  -H "Content-Type: application/json" \
  -d '{"input": "How do I reset my password?"}'

This is how other apps would integrate with your bot — any frontend (mobile app, Slack bot, widget) can call this endpoint.

Monitor in the Dashboard

Go to your Kernova Cloud dashboard:

  • Projects → See your deployed project
  • Monitoring → Check deployment health, success rate
  • Deployments → View build logs and history

Make a Change and Redeploy

Let’s add a new article to the knowledge base. Open agents/tools/search-kb.ts and add to the array:

{
  id: 'kb-007',
  title: 'Mobile app',
  content: 'Our mobile app is available on iOS and Android. Download from the App Store or Google Play. Sign in with your existing account credentials.',
  tags: ['mobile', 'app', 'ios', 'android'],
},

Now redeploy:

Terminal window
kernova deploy

Within seconds, your production bot can answer questions about the mobile app. Zero downtime — the new version replaces the old one seamlessly.

What You Built

Congratulations! You’ve built and deployed a production support bot with:

✅ An AI agent with custom instructions
✅ A knowledge base search tool
✅ A ticket creation tool
✅ Streaming chat UI
✅ Conversation memory
✅ One-command deployment
✅ Live at a public URL

What’s Next

From here, you could:

  • Add more agents — Create agents/billing.agent.ts for billing-specific questions
  • Connect a real database — Replace the hardcoded KB with a vector database
  • Add authentication — Protect the chat page so only logged-in users can access it
  • Embed in your app — Use the API endpoint from your existing website’s chat widget
  • Set up custom domain — Point support.yourcompany.com to your deployment

Full Project Recap

Here’s your final project structure:

support-bot/
├── app/
│   ├── layout.tsx           # KernovaProvider wrapper
│   ├── page.tsx             # Landing page
│   ├── globals.css          # Chat styles
│   └── chat/
│       └── page.tsx         # Chat UI with useAgentStream
├── agents/
│   ├── support.agent.ts     # Support agent definition
│   └── tools/
│       ├── search-kb.ts     # Knowledge base search
│       └── create-ticket.ts # Ticket creation
├── memory/
│   └── conversations.ts     # Memory schema
├── kernova.config.ts
├── next.config.ts
├── .env.local
└── package.json

Commands used:

Terminal window
npx create-kernova-app support-bot   # Create
kernova dev                           # Develop
kernova login --key kn_live_xxx       # Authenticate
kernova deploy                        # Ship

That’s the Kernova developer experience: define agents, build a UI, deploy with one command.