# Scraper MCP server

> Reduces LLM token usage by converting web content to markdown and applying server-side CSS filtering for optimized web scraping.

- Repository: https://github.com/cotdp/scraper-mcp
- Page: https://hostedmcp.org/servers/cotdp/scraper-mcp
- Transport: —
- Stars: 8
- License: MIT

## Environment variables

| Name | Required | Secret | Description |
|---|---|---|---|
| `ENABLE_PROMPTS` | no | no |  |
| `ENABLE_RESOURCES` | no | no |  |
| `OPENROUTER_API_KEY` | no | no |  |
| `PERPLEXITY_API_KEY` | no | no |  |
| `PERPLEXITY_PROVIDER` | no | no |  |
| `PLAYWRIGHT_DISABLE_GPU` | no | no |  |
| `PLAYWRIGHT_MAX_CONTEXTS` | no | no |  |
| `PLAYWRIGHT_TIMEOUT` | no | no |  |
| `SCRAPEOPS_API_KEY` | no | no |  |
| `SCRAPEOPS_RENDER_JS` | no | no |  |

## Tools (6)

- `scrape_url`
- `scrape_url_html`
- `scrape_url_text`
- `scrape_extract_links`
- `perplexity`
- `perplexity_reason`

## FAQ

### How does this server reduce LLM token usage?

It uses server-side HTML filtering, markdown conversion, and CSS selector targeting to strip unnecessary data.

### Can I scrape websites that require JavaScript to load content?

Yes, you can enable optional Playwright-based rendering for SPAs and dynamic content using the render_js parameter.