# Scrapling MCP Server

**Category:** 🔎 Search & Data Extraction  
**Repository:** https://github.com/D4Vinci/Scrapling  
**GitHub Stars:** 81460  
**Views:** 3  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/scrapling-mcp-server

## Description
Web scraping with stealth HTTP, real browsers, and Cloudflare bypass. CSS selectors supported.

## Claude Desktop Quick Installation
Install path detected from listing signals. Uses `uvx` (confidence: high):

```json
"mcpServers": {
  "scrapling-mcp-server": {
    "command": "uvx",
    "args": ["scrapling"]
  }
}
```

## Documentation

## What Scrapling MCP Server MCP server does

Scrapling MCP Server MCP server makes Scrapling’s web-scraping capabilities available through the Model Context Protocol. The underlying framework covers use cases ranging from a single page request to larger crawls. It can retrieve pages with ordinary or stealth-oriented fetchers, use real browsers for dynamic sites, and parse returned content with CSS selectors.

The parser can save information about page elements and later use adaptive matching to relocate those elements after a site changes its structure. This is relevant for agents that need to extract repeated fields from pages whose markup may evolve over time. The framework also includes spiders for concurrent, multi-session crawling, along with pause and resume support, proxy rotation, streaming, and real-time crawl statistics.

As a Search & Data Extraction listing, it fits workflows where an agent must read public web content rather than interact with a structured service API. The supplied material identifies MCP support but does not document the exact MCP operations exposed to clients.

## How it works

A fetcher obtains a web page, optionally using stealth behavior or a browser-backed approach. The resulting response can then be queried with CSS selectors. The README examples show selectors being used to identify collections such as products and to retrieve text from nested elements.

For larger jobs, Scrapling’s spider model starts from one or more URLs and processes responses asynchronously. Spider behavior can include concurrent sessions, automatic proxy rotation, adaptive crawl speed, and backoff when a site begins blocking requests. These capabilities belong to the Scrapling framework; the supplied material does not explain which of them are directly selectable through the MCP layer.

The project describes fetchers that can bypass anti-bot systems such as Cloudflare Turnstile. Browser-based fetching and headless operation are also shown in the examples, but no MCP-specific parameters are provided.

## Setup and configuration

The source material links to Scrapling’s MCP documentation and identifies this repository as an MCP server, but it does not include installation commands or a client configuration example. It also does not list required environment variables, command-line arguments, transport settings, or authentication requirements.

The repository is published as a Python package and also has a Docker image reference in its project metadata. Those references establish available distribution channels for Scrapling, but they do not, by themselves, define the setup procedure for the MCP server. Consult the repository’s MCP documentation before deployment.

## Tools and capabilities

The documented capabilities include:

- Fetching pages through standard, stealth, asynchronous, and dynamic/browser-oriented fetchers.
- Running headless browser requests and waiting for network activity to become idle.
- Selecting page content with CSS selectors.
- Relocating saved selections when page structure changes through adaptive parsing.
- Building concurrent, multi-session crawls with spiders.
- Pausing and resuming crawls, rotating proxies, and adjusting crawl speed in response to site behavior.
- Producing streaming output and real-time crawl statistics.

Scrapling MCP Server MCP server may expose some of these functions to an agent, but the provided excerpt does not enumerate MCP tool names or their input and output schemas.

## Limitations and notes

The material does not state which websites can be accessed successfully, guarantee that anti-bot protections will be bypassed, or describe compliance requirements for scraping. Cloudflare Turnstile bypass is presented as a framework capability, not as a guarantee for every target or configuration.

No supported MCP client is named in the supplied information. Claude Desktop, Cursor, Windsurf, and Cline should therefore be treated as unconfirmed for this listing. The excerpt also provides no license declaration, required environment variables, rate-limit guidance, or authentication details.

For an agent workflow, verify the MCP documentation for exposed operations, browser dependencies, proxy settings, and response-size behavior before choosing Scrapling MCP Server MCP server.

_Full upstream README: https://allmcps.com/mcp/scrapling-mcp-server/readme_