explaingit

jcodesmore/ai-website-cloner-template

14,708TypeScript

TLDR

AI Website Cloner Template is a starter project that helps an AI coding assistant rebuild any public website as a fresh, modern Next.js codebase.

Mindmap

A visual breakdown will appear here once this repo is fully enriched.

In plain English

AI Website Cloner Template is a starter project that helps an AI coding assistant rebuild any public website as a fresh, modern Next.js codebase. You clone the template, point it at one or more URLs, and run a slash command called /clone-website inside an AI agent. The agent then visits the target site, takes screenshots, extracts the design details, writes specifications for each piece of the page, and dispatches separate worker agents to rebuild every section in parallel. The authors recommend pairing it with Claude Code running the Opus 4.7 model, but the README lists more than a dozen supported assistants, including Codex CLI, OpenCode, GitHub Copilot, Cursor, Windsurf, Gemini CLI, Cline, Roo Code, Continue, Amazon Q, Augment Code, and Aider. A file called AGENTS.md holds the master set of project instructions, and small platform-specific copies for Claude and Gemini are generated from it by a sync script so that whichever agent you use picks up the same rules. The pipeline the README describes runs in five phases. First the agent does reconnaissance, taking screenshots and probing the site by scrolling, clicking, and resizing. Then it lays the foundation, copying over fonts, colours, global styles, and downloaded assets. After that it writes detailed component specs that record exact CSS values, interaction states, and responsive behaviour. Builder agents then work in separate git worktrees, each handling one section, before everything is merged together and a visual diff against the original is run as a quality check. The generated project is built on Next.js 16 with the App Router, React 19, TypeScript in strict mode, Tailwind CSS v4 with oklch design tokens, shadcn/ui components on top of Radix primitives, and Lucide icons that get swapped out for SVGs extracted from the target. The README is upfront about what the tool should not be used for, including phishing, impersonating brands, passing off someone else's design as your own, or scraping sites whose terms forbid it. Suggested legitimate uses are migrating off WordPress or Webflow, recovering a site whose source code is lost, and studying production sites by examining real reconstructed code. The project is MIT licensed.

Open on GitHub → Explain another repo

Generated 2026-05-21 · Model: sonnet-4-6 · Verify against the repo before relying on details.