M
MCP Directory CURATED DIRECTORY
911 MCP servers indexed
中文
← Back to directory
S

Scrapling

Official
Web scraping MCP server with stealth requests, real browsers, and adaptive selectors
How to install ↓ Compare GitHub source repository ↗
Category
Other #131 of 230
Stars
★ 88.5k Very popular
Transport
stdio (local process)
Runtime
Python 3.10+ · Docker
Credentials
No credential needed
License
BSD-3-Clause
Last commit
52FMRS · D

An official MCP server included in the upstream Scrapling repository. Version 0.4.12 uses stdio transport and targets AI-assisted web scraping and data extraction. It is distributed through PyPI and OCI packages; the source provides no named tools or client compatibility list.

Strongest · Maintenance 13/20 Weakest · Security and permissions 7/20

Reliability
9/20
Security and permissions
7/20
Maintenance
13/20
Documentation
12/20
Setup experience
11/20
Why each score
Reliability 9/20
The manifest declares a version, PyPI and OCI packages, and stdio transport, while the README describes MCP functionality, making a normal startup and handshake plausible. However, there is no supplied tool list, tool-parameter schema, server implementation, committed MCP-path tests, or CI evidence covering the key server paths. Browser dependencies, optional extras, and external network access are also not bounded by documented error behavior, so points are deducted conservatively.
Security and permissions 7/20
The server accesses external websites and supports stealth requests, proxy rotation, browser automation, and Cloudflare bypass, potentially giving it broad network reach. The supplied material does not explain credential handling, proxy-secret protection, domain scoping, user confirmation, logging/data retention, or permission isolation. No malware, credential theft, or default delete/pay/deploy behavior is evidenced, but the powerful network capabilities and missing controls warrant a low score.
Maintenance 13/20
The project is not archived and provides a BSD-3-Clause license, identifiable repository ownership, version 0.4.12, PyPI and OCI release paths, a test-workflow reference, and contribution guidance. The material does not establish commit cadence, dependency-update practice, vulnerability-response procedures, or actual CI results. Issue count and star count are discovery signals only and add no points, so full marks are not justified.
Documentation 12/20
The README covers features, installation, optional dependencies, browser setup, Docker, usage examples, robots.txt behavior, legal limitations, and a link to MCP documentation. It does not provide the MCP tool names, parameter and return schemas, authentication model, resource/concurrency/cost limits, troubleshooting, or complete mainstream-client configuration. Several capability claims are project assertions rather than independently reviewable evidence, so points are deducted.
Setup experience 11/20
The manifest supplies PyPI and OCI installation options, a uvx runtime hint, a fixed mcp argument, and stdio transport; the README documents Python requirements, extras, browser installation, and Docker. It lacks a complete configuration example for a mainstream MCP client, authentication guidance, platform compatibility details, and troubleshooting for startup failures. Browser dependencies may also make installation more complex, so the score remains moderate.

Static review · not runListed 2026-07-29

Read the FMRS scoring method →

Fit and risk

What it can accessUses the networkControls a browser

Best for

  • Users who need AI-assisted web data extraction
  • Python web-scraping and automation developers
  • Projects requiring browsers, sessions, proxy rotation, or adaptive selectors

Not for

  • Local data processing that does not access external websites
  • Workflows limited to static files or cloud-storage operations
  • Use cases without lawful scraping permission or without adherence to website terms and robots.txt

Required permissions

  • Access to target websites and permission to send HTTP requests or launch browsers
  • Depending on proxy, session, or browser features, corresponding network and local runtime permissions may be required
  • The source does not specify account authentication or particular API credentials

Risks and side effects

  • Scraping may violate website terms of service, privacy regulations, or applicable data-collection laws
  • Stealth requests, browser automation, and Cloudflare bypass may trigger website security controls
  • Proxies, cookies, session state, and scraped content may contain sensitive information
  • High-concurrency or large-scale crawling may impose load on target websites

Setup

Before you start

Runtime：Python 3.10+ · Docker

Run the PyPI package scrapling with uvx and pass the fixed argument mcp. An OCI package, ghcr.io/d4vinci/scrapling, is also available with the mcp argument. The server communicates over stdio.

Check that it works

The Scrapling MCP server's tools should appear in your client's tool list; asking the AI to fetch a simple page and return its content confirms the connection works.

Troubleshooting

  1. Confirm that scrapling and the MCP-related dependencies are installed
  2. Confirm that the client supports stdio MCP servers
  3. When using browser or fetching features, install browser dependencies with scrapling install
  4. Check that the target URL and CSS or XPath selectors are valid
  5. If requests are blocked, check proxies, sessions, stealth settings, robots.txt, and target-site restrictions

Things to try

Once connected, you can ask your AI assistant things like:

  • Fetch https://example.com and extract the main page content
  • Use stealth mode to scrape https://nopecha.com/demo/cloudflare and return the page title
  • Take a screenshot of this page and summarize what's on it

Use cases

Extract structured data from web pages
Scrape JavaScript-rendered websites
Locate content with CSS or XPath selectors
Handle websites with anti-bot protections
Run tasks ranging from single-page requests to concurrent crawls

Overview

Scrapling MCP Server supports AI-assisted web scraping and data extraction with stealth HTTP requests, real browsers, Playwright, Cloudflare Turnstile bypass, and CSS selectors. The underlying framework supports tasks ranging from single requests to full-scale crawls, adaptive element tracking, XPath, and session management.

Similar servers

BioMCP 75 · B

One binary. One grammar. Evidence from the biomedical sources you already trust.

★ 659 · Tools 11 Compare with this →
Compare with other MCP servers

Source revision fd531313a499 Data synced 2026-10-19 Read the FMRS scoring method