# {{BIN_NAME}}

{{DESCRIPTION}}

A Playwright-based MCP server for scraping JavaScript-rendered pages.

## Prerequisites

```bash
npx playwright install chromium
```

## Install

```bash
# Claude Code
claude mcp add {{PROJECT_NAME}} -- npx {{BIN_NAME}}

# Claude Desktop — add to claude_desktop_config.json
{
  "mcpServers": {
    "{{PROJECT_NAME}}": {
      "command": "npx",
      "args": ["{{BIN_NAME}}"]
    }
  }
}
```

## Tools

| Tool | Description |
|------|-------------|
| `scrape_page` | Load page in browser, extract text + metadata |
| `screenshot` | Take viewport or full-page screenshot |

## Architecture

```
src/
└── index.ts    # MCP server + Playwright browser management
```

Uses a shared browser instance for performance. Browser context is created per-request for isolation.

## Development

```bash
npm install
npx playwright install chromium
npm run dev    # Start with hot reload
npm run build  # Compile TypeScript
npm publish    # Publish to npm
```

## License

MIT

---

Created with [mcp-quick](https://github.com/yifanyifan897645/mcp-quick)