Cube

Create fully-functional web applications by chatting with an AI.

Cube is a generative AI application that allows you to describe a web page or component in plain English and watch as it gets built and deployed in real-time. It leverages a sophisticated multi-agent system to interpret requests, write code, and run it in a secure cloud environment.

✨ Key Features

• From Idea to Live App, Instantly: Describe the web application you want, and our AI will build and deploy it for you in real-time.
• Live, Interactive Previews: See your generated website come to life in a real-time preview and interact with it as you build.
• Live & Secure Sandboxes (Powered by E2B): Each project is built in a dedicated cloud sandbox, providing a secure, instantly-deployed environment and a shareable live URL.
• Full Code Access & Transparency: Get complete access to the generated source code, with the ability to view and download it at any time.

⚙️ How It Works

Cube's magic lies in its robust, event-driven architecture that coordinates multiple services to turn a prompt into a running web application.

1. User Prompt: The user submits a request through the Next.js web interface.
2. Orchestration Layer: An Inngest function picks up the request and begins a durable, long-running workflow.
3. Secure Sandbox Creation: Inngest spins up a secure, isolated E2B cloud sandbox for the AI to work in.
4. AI Code Generation: A Google Gemini-powered AI agent is invoked inside the sandbox. It has access to a file system, a terminal, and other tools necessary to build a web application.
5. Real-Time Deployment: As the agent works, it runs a development server inside the sandbox, which is exposed via a secure, public URL for the user to preview.
6. Saving the Result: Once the build is complete, the final code is saved to the database via Prisma, and the user is presented with the final result.

🚀 Tech Stack

• Framework: Next.js
• Language: TypeScript
• AI Orchestration: Inngest
• AI Sandboxing: E2B
• AI Model: Google Gemini
• Database: PostgreSQL via Prisma
• Authentication: Clerk
• UI: Tailwind CSS & Shadcn UI
• Deployment: Vercel

🏁 Getting Started

Follow these instructions to set up and run the project on your local machine.

Prerequisites

• Node.js (v20.x or later)
• Bun
• Git
• PostgreSQL

1. Clone the Repository

git clone https://github.com/your-username/cube.git
cd cube

2. Install Dependencies

bun install

3. Set Up Environment Variables

Create a .env.local file in the root of the project by copying the example file:

cp .env.example .env.local

Now, fill in the required values in .env.local:

# Prisma
DATABASE_URL="postgresql://user:password@host:port/database"

# Clerk
NEXT_PUBLIC_CLERK_PUBLISHABLE_KEY=
CLERK_SECRET_KEY=

# Inngest
INNGEST_EVENT_KEY=
INNGEST_SIGNING_KEY=

# E2B
E2B_API_KEY=

# Google Gemini
GOOGLE_API_KEY=

4. Set Up the Database

Run the Prisma migrations to set up your database schema.

bunx prisma migrate dev

5. Run the Development Server

Now you can start the development server:

bun run dev

Open http://localhost:3000 with your browser to see the result.

🤝 Contributing

Pull requests are welcome! Feel free to fork this repo, open issues, and suggest enhancements.

📄 License

This project is licensed under the MIT License. See the LICENSE file for details.

📬 Author

Gaurav Vashisth
Github | Email: gauravvashisth365@gmail.com

🫂 Contributors

Aryan Shandilya
Deepansh Sharwan