Create a Project

Using npx (recommended)

npx create-kernova-app my-app
cd my-app
kernova dev

This scaffolds a complete project with a Next.js frontend and agent backend.

Options

# Default template (1 agent + chat UI)
npx create-kernova-app my-app

# Multi-agent template (support + researcher agents)
npx create-kernova-app my-app --template multi-agent

# Agents only (no frontend)
npx create-kernova-app my-app --no-frontend

# Skip npm install
npx create-kernova-app my-app --no-install

Using the CLI

If you already have @kernova/cli installed globally:

kernova create my-app
kernova create my-app --template multi-agent
kernova create my-app --no-frontend

What You Get

After running create-kernova-app, you’ll have:

my-app/
├── app/
│   ├── layout.tsx         # Wraps your app with KernovaProvider
│   ├── page.tsx           # Landing page
│   ├── globals.css        # Starter styles
│   └── chat/
│       └── page.tsx       # Chat page using useAgentStream
├── agents/
│   ├── assistant.agent.ts # Your first agent
│   └── tools/
│       └── web-search.ts  # Example tool
├── memory/
│   └── conversations.ts   # Memory schema
├── kernova.config.ts      # Project configuration
├── next.config.ts         # Next.js rewrites (auto-configured)
├── .env.local             # Your API keys go here
├── package.json
└── tsconfig.json

Configuration

The kernova.config.ts file defines your project settings:

import { defineConfig } from '@kernova/framework/config';

export default defineConfig({
  name: 'my-app',
  model: {
    provider: 'openai',
    model: 'gpt-4o',
  },
  memory: {
    provider: 'local',
    embeddingModel: 'text-embedding-3-small',
  },
  gateway: {
    openaiCompat: true,
    cors: ['*'],
    rateLimit: { requests: 100, window: '1m' },
  },
  app: {
    framework: 'next',
    port: 3000,
  },
});

Environment Variables

Add your API keys to .env.local:

OPENAI_API_KEY=sk-...

That’s it. Run kernova dev and your agent is ready to talk.

Development Commands

Command	What It Does
`kernova dev`	Start frontend + agents (one URL)
`kernova build`	Build for production
`kernova start`	Run production server locally
`kernova deploy`	Deploy to Kernova Cloud