Lovable Clone - AI-Powered Web App Builder

A real-time AI-powered web application builder inspired by lovable.dev.

️Architecture

This project demonstrates how to build a simple agent using sandboxed environments, MCP servers, and BAML. It's hosted on beam.cloud.

To learn about the architecture in detail, read the full post on our blog.

The application consists of three main components:

1. Model Client - Communication with LLM, based on BAML
2. Sandbox Environment - Runs the React preview in a compute sandbox on Beam.
3. WS-Based Agent - Streams edit requests from the user to the agent

Prerequisites

• Python 3.12+
• Node.js 20+
• OpenAI API key
• Beam account for hosting (sandboxes, frontend, mcp server)

Installation

1. Clone and install

   git clone https://github.com/beam-cloud/lovable-clone
   cd lovable-clone
   pip install -r requirements.txt

2. Install frontend dependencies

   cd frontend
   npm install

3. Set up secrets

You can grab your OpenAI API key here.

beam secret create OPENAI_API_KEY 'your-openai-api-key'

4. Add .env

Create an .env file in /frontend

You’ll need a Beam token and the websocket URL, printed in your shell after running the command below.

VITE_BEAM_WS_URL=your-agent-ws-url # URL printed in your shell after running `beam serve` - fill this in after running the command below
VITE_BEAM_TOKEN=your-beam-token # Get Beam token here: https://platform.beam.cloud/settings/api-keys

Usage

Start the Agent

The agent handles real-time communication with the client/frontend:

beam serve src/agent.py:handler

Paste the websocket URL printed in your shell into your .env file above

Run the Frontend

In a new terminal window, run:

cd frontend
npm run dev

BAML / Prompts

Prompts are defined in baml_src/build.baml:

• EditCode Function - Main function for code generation
• CodeChanges Schema - Defines the structure of AI responses
• Test Cases - Validate prompt behavior

If you want to change the prompt, edit baml_src/build.baml and run make generate to regenerate the BAML clients

Sandbox Environment

The sandbox environment is managed in src/tools.py:

• Node.js 20 base image
• React + Vite + shadcn/ui template
• Other deps: React Router, Recharts, TanStack Query, etc.