Skip to content

Cowpy

Cowpy is a cat that turns any live website into a production-ready Next.js 15 app from a single URL.

Core trick

Records browser sessions and HAR traffic, then uses frame-by-frame video analysis to reproduce animations in a generated Next.js app.

Tags

Description

Cowpy is a cat that turns any live website into a production-ready Next.js 15 app from a single URL. Unlike screenshot-to-code tools that ask an LLM to guess markup from a static image, Cowpy runs an 8-step agentic pipeline: it crawls the sitemap, records browser sessions as video, captures HAR network traffic, detects animations frame-by-frame via video analysis, and generates a full Next.js app The result is a real, buildable app with routes, animations, hover states, and API integrations — not a static HTML approximation. The entire pipeline is orchestrated through Claude Code skills and subagents, coordinating Claude Opus 4.6, Sonnet 4.5, Haiku 4.5, Kimi K2.5 and Flux across specialized roles.
DiasKhalniyasov/full-clone-site2code★ 20 forksPythonCreated Feb 11, 2026 · last push Feb 16, 2026

README excerpt

# Cowpy - Site2Code

Clone any website into a working Next.js 15 app — from a single URL.

```
/copy-website https://example.com
```

Site2Code is a [Claude Code](https://docs.anthropic.com/en/docs/claude-code) skill that crawls a website, records interactions, analyzes visual design and animations, and generates a pixel-perfect Next.js 15 application with Tailwind CSS v4 and Framer Motion.

## Setup

### Prerequisites

- Python 3.11+
- Node.js 18+
- [Claude Code](https://docs.anthropic.com/en/docs/claude-code) CLI


### Set key in .env file:
Kimi K2.5 is required for video analysis (for animations). used as default.
Flux 2 Klein is required for image generation.
Both are optional.
Gemini 3 Flash is optional for video analysis (faster than Kimi). use `--video-provider gemini` flag.


### Installation

```bash
# Clone the repo
git clone <repo-url> && cd video-to-code

# Create venv and install Python dependencies
python3 -m venv .venv
.venv/bin/pip install -r requirements.txt

# Install Playwright browsers
.venv/bin/python -m playwright install chromium

# Configure environment
cp .env.example .env
# Edit .env and fill in your API keys (see Environment Variables below)
```

The `.claude/` directory contains the skill definitions — Claude Code picks them up automatically.

## Quick Start

```bash
# Clone a website (runs the full 10-step pipeline)
/copy-website https://example.com

# View the result
cd output/generated && npm run dev
```

## Available Models

The pipeline uses m

Team