This project provides a template for reverse-engineering and pixel-perfectly cloning any website using AI coding agents. It targets developers and designers seeking an automated way to replicate web designs, offering a significant time-saving benefit by abstracting complex inspection and reconstruction tasks.

How It Works

The system employs a multi-phase AI pipeline starting with reconnaissance (screenshots, design token extraction, interaction sweeps). It then establishes a foundation (fonts, colors, globals) and generates detailed component specifications with exact CSS values and behaviors. Finally, parallel builder agents, each operating in a dedicated git worktree, reconstruct sections based on these specs, followed by assembly and visual diffing against the original site. This approach ensures high fidelity by providing builders with precise, inline specifications, eliminating guesswork.

Quick Start & Requirements

• Install dependencies: npm install
• Prerequisites: Node.js 20+, an AI coding agent (Claude Code Opus 4.6 recommended).
• Run: Execute /clone-website <target-url> via your chosen AI agent (e.g., claude --chrome).
• Docs: AGENTS.md for agent instructions.

Highlighted Details

• Supports a wide array of AI coding agents, including Claude Code, Gemini CLI, GitHub Copilot, and Amazon Q.
• Leverages Next.js 16 (App Router), React 19, and Tailwind CSS v4 with oklch design tokens for modern web development.
• Generates detailed component specifications with exact getComputedStyle() values, interaction models, and responsive breakpoints.
• Utilizes parallel builder agents in separate git worktrees for efficient, section-specific reconstruction.

Maintenance & Community

The README does not provide specific details on maintainers, community channels (like Discord/Slack), or a public roadmap. Configuration for different agents is managed via AGENTS.md and platform-specific sync scripts.

Licensing & Compatibility

Licensed under MIT, permitting commercial use and integration into closed-source projects without significant restrictions.

Limitations & Caveats

The effectiveness is highly dependent on the chosen AI coding agent's capabilities, with Claude Code Opus 4.6 specifically recommended for optimal results. The project relies on external AI agents, and setup may require familiarity with specific agent CLIs or configurations.