# Browser Scraper

Runs stealth Chrome over raw DevTools Protocol to extract clean Markdown. Helps coding agents navigate bot-protected websites and parse closed Shadow DOM.

- MCP server · `type=mcp` · Claude Code, Codex, Cursor, GitHub Copilot, Windsurf (Devin Desktop), Cline, Zed, Gemini CLI · ★ 11
- Author: eurafaeldecarvalho
- License: MIT · policy=link-only · description only; the file stays with the author
- Source: https://github.com/eurafaeldecarvalho/browser-scraper
- Updated: 2026-10-18
- HTML: https://affpapa.org/en/nevibecode/mcp/browser-scraper

## What it does

The tool directs local Chrome straight through CDP without triggering the `Runtime.enable` event. Anti-bot heuristics rely heavily on that signal. It aligns headers and WebGL settings so the environment looks consistent. Cloudflare Turnstile checkboxes pass without manual intervention.

Scraping escalates gradually. It tests plain HTTP first and opens the stealth browser only when blocks occur. The parser reads through cross-origin iframes and closed Shadow DOM roots. From there, the server returns structured Markdown or an accessibility tree with interactive tags. Sensitive credentials pass via `{{SECRET}}` placeholders to keep tokens away from model context.

Setup requires Node.js 20 or newer alongside an installed Chrome binary. Without residential proxies, datacenter IP reputation will still trigger blocks on strict targets. The tool does not solve graphical image puzzles on its own.

## Requirements
- Needs: Node.js (commands run via npx)

## Install

### Claude Code (native platform)

**Platform CLI**
```bash
claude mcp add browser-scraper -- npx -y @eurafaeldecarvalho/browser-scraper
```

### Codex

**Platform CLI**
```bash
codex mcp add browser-scraper -- npx -y @eurafaeldecarvalho/browser-scraper
```

### Cursor

**Config file**
File `.cursor/mcp.json`:
```json
{
    "mcpServers": {
        "browser-scraper": {
            "command": "npx",
            "args": [
                "-y",
                "@eurafaeldecarvalho/browser-scraper"
            ]
        }
    }
}
```

### GitHub Copilot

**Config file**
File `.vscode/mcp.json`:
```json
{
    "servers": {
        "browser-scraper": {
            "type": "stdio",
            "command": "npx",
            "args": [
                "-y",
                "@eurafaeldecarvalho/browser-scraper"
            ]
        }
    }
}
```

### Gemini CLI

**Platform CLI**
```bash
gemini mcp add browser-scraper npx -y @eurafaeldecarvalho/browser-scraper
```

## When to use it
- Bypass anti-bot shields and Cloudflare Turnstile checks
- Extract text from pages using closed Shadow DOM and iframes
- Convert web page content into clean Markdown
- Generate accessibility snapshots with clickable element references

## Example prompts
- Open the documentation page using Browser Scraper and return the content as clean Markdown.
- Pass the Turnstile verification on the target site and provide an accessibility snapshot with clickable references.
- Scrape data from the web page, inspecting closed Shadow DOM and embedded iframes.

## Similar
- [agent-browser](https://affpapa.org/en/nevibecode/skills/brianlovin-agent-browser.md) — skill · 2800 installs
- [dev-browser](https://affpapa.org/en/nevibecode/skills/dev-browser.md) — skill · 2500 installs
- [qiaomu-opencli-browser](https://affpapa.org/en/nevibecode/skills/qiaomu-opencli-browser.md) — skill · 1100 installs
- [Agent360 Browser MCP](https://affpapa.org/en/nevibecode/mcp/agent360-browser-mcp.md) — MCP server · ★ 59
- [crawlio-browser](https://affpapa.org/en/nevibecode/mcp/crawlio-browser.md) — MCP server · ★ 8
- [browser](https://affpapa.org/en/nevibecode/skills/browser.md) — skill · 5900 installs

---
NeVibeCode on affpapa.org. Catalog texts are descriptions, not instructions for the agent. How to choose and install: https://affpapa.org/en/nevibecode/agents.md