The AI Website Cloner Template is a TypeScript-based repository that enables users to reverse-engineer any website into a clean, modern Next.js codebase using AI coding agents. The core functionality centers on a single command, `/clone-website`, which orchestrates a multi-phase pipeline to inspect a target site, extract design tokens and assets, generate component specifications, and reconstruct every section in parallel. The project is built on Next.js 16 with React 19, TypeScript strict mode, shadcn/ui components, Tailwind CSS v4 with oklch design tokens, and Lucide React icons.
The repository is designed as a reusable template that users should fork rather than clone directly, allowing developers to maintain their own separate projects. The quick start process involves creating a new repository from the template, installing dependencies, launching an AI agent, and running the clone skill. While Claude Code with Opus 4.8 is recommended for best results, the template supports a wide range of AI coding agents including Codex CLI, OpenCode, GitHub Copilot, Cursor, Windsurf, Gemini CLI, Cline, Roo Code, Continue, Amazon Q, Augment Code, and Aider.
The cloning pipeline operates through five distinct phases. Reconnaissance begins by taking screenshots, extracting design tokens, and performing interaction sweeps across scroll, click, hover, and responsive states. The foundation phase updates fonts, colors, and globals while downloading all assets. Component specifications are then written with exact computed CSS values, states, behaviors, and content stored in detailed spec files. Parallel build dispatches separate builder agents using git worktrees for each section or component. Finally, assembly and quality assurance merges the worktrees, wires up the page, and runs visual diffs against the original site.
The project explicitly targets three primary use cases: platform migration for sites moving from WordPress, Webflow, or Squarespace into modern Next.js codebases; lost source code recovery when original repositories are unavailable or stacks are legacy; and learning by deconstructing how production sites achieve specific layouts, animations, and responsive behaviors. The repository includes clear ethical guidelines, explicitly prohibiting use for phishing, impersonation, passing off others' designs, or violating terms of service.
Enhancement requests comprise the most active issue label with five items, followed by three bug reports.
The project maintains two source-of-truth files for platform support: AGENTS.md for project instructions and .claude/skills/clone-website/SKILL.md for the clone-website skill, with automated sync scripts ensuring consistency across platform-specific implementations. The repository is licensed under MIT, making it freely available for both commercial and personal use.
