A modern AI-powered coding assistant that generates and deploys code to a secure sandbox in real-time through natural language conversations. Built with Cerebrium and E2B, this system provides an intuitive interface for turning your ideas into working code in no time.
- Real-time Code Generation: Watch as code is generated instantly as you describe your needs
- Live Preview: See your code running in a sandboxed environment
- File Tree Navigation: Browse generated files with ease
- Describe Your Needs: Start a conversation with the agent about what you want to build
- Watch Code Generation: See code being written in real-time with syntax highlighting
- Preview Instantly: Click the preview link to see your code running live
- Frontend: Next.js 14, TypeScript, Tailwind CSS, shadcn/ui (Access the frontend code, here)
- Backend: FastAPI, WebSocket for real-time communication
- LLM: Qwen 2.5 7B models for code generation running on Cerebrium cloud infrastructure
- Preview: Sandboxed environment for secure code execution running on E2B
- LLMs are capable of hallucinating. As we're using a large language model with half-precision, it may generate code that doesn't work as expected. Please review the generated code before running it in a production environment.
# Clone the repository
git clone https://github.com/CerebriumAI/examples.git
# Deploy AI coding agent example to cerebrium
cd examples/8-application-demos/4-ai-coding-agent
cerebrium deploy -y
# Connect to your deployed socket endpoint and start sending messages