proxyclaw-mcp-py
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@proxyclaw-mcp-pyScrape Amazon search results for 'laptop' and extract prices"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
ProxyClaw MCP Server (Python)
Model Context Protocol server for ProxyClaw. It gives AI agents controlled residential-proxy web access, optional browser rendering, and structured extraction workflows for legitimate testing, research, and automation use cases.
⚡ Python vs Node.js — Which One?
We ship two MCP servers. Choose based on what you need:
This repo (Python) | ||
What it does | Browser rendering + structured extraction | Proxy routing + fetch |
Best for | JS-rendered pages and structured extraction workflows | Simple fetches, geo-targeting |
Networking | Optional TLS-compatible client libraries + Playwright rendering | Standard proxy-backed fetch |
Install |
|
|
Tools | 6 (+ stealth fetch, render, scrape, extract) | 4 (fetch, check_ip, list_countries, rotate) |
→ Use Python if you need JavaScript rendering or structured data extraction from supported sites.
→ Use Node.js if you just need to route requests through residential IPs.
Both use the same proxy network — just different levels of power.
Related MCP server: krwl3r
Install
# Core proxy workflow
pip install proxyclaw-mcp-server
# Optional compatible HTTP clients
pip install proxyclaw-mcp-server[stealth]
# With browser rendering
pip install proxyclaw-mcp-server[all]
# Or via uv (Claude Desktop recommended)
uv tool install proxyclaw-mcp-server[all]Claude Desktop Config
Add to claude_desktop_config.json:
{
"mcpServers": {
"proxyclaw": {
"command": "uvx",
"args": ["proxyclaw-mcp-server[all]"],
"env": {
"IPLOOP_API_KEY": "your-api-key"
}
}
}
}Tools
Tool | What it does |
| Fetch URL through ProxyClaw with optional compatible HTTP clients + retry |
| Headless Chromium render for JavaScript apps and dynamic pages |
| Smart cascade: direct → stealth → render → search fallback |
| Structured data from 60+ sites (eBay, Amazon, LinkedIn, Google, Nasdaq, etc.) |
| Verify exit IP and geo-location |
| List 195+ available countries |
Quick Start
Get your API key at iploop.io/signup.
Then ask Claude:
"Fetch https://example.com through a German IP with stealth headers"
"Render https://spa-site.com and extract the product title"
"Scrape eBay search results for 'laptop' and return structured data"
Features
Residential proxy network — 175M+ real IPs across 195+ countries
Compatible HTTP clients — optional curl_cffi / tls_client integrations for more browser-like request behavior
Headless rendering — Playwright rendering for JavaScript-heavy pages
Structured extraction — 60+ site-specific parsers (eBay, Amazon, LinkedIn, Twitter, Nasdaq, etc.)
Smart cascade — tries direct fetch first, escalates to stealth → render → search fallback
Sticky sessions — keep same IP across requests
Cookie persistence — cookies saved between requests
Auto-retry with rotation — configurable retry behavior for transient proxy/network failures
Security & Trust
Source code: https://github.com/Iploop/proxyclaw-mcp-py
Security page: https://proxyclaw.ai/security.html
The package is a thin MCP adapter around the public
iploop-sdk; it does not include install hooks, shell execution, credential exfiltration, crypto mining, or obfuscated payloads.Runtime credential access is limited to
IPLOOP_API_KEY, used only to authenticate proxy/API requests.
Links
Website: proxyclaw.ai
Security: proxyclaw.ai/security.html
Dashboard: iploop.io/dashboard
Python MCP source: Iploop/proxyclaw-mcp-py
Python SDK: iploop on PyPI
Node.js SDK: iploop on npm
License
MIT
This server cannot be deployed
Maintenance
Related MCP Connectors
Reliable web access for AI agents: smart HTTP, rotating proxies, and full-browser rendering.
MCP server connecting AI agents to non-custodial staking data across 130+ networks.
Scrape, crawl and search the web for AI agents via MCP.
Zenrows MCP server — Fetch, Extract, Batch, and Browser Sessions for AI coding assistants
Related MCP Servers
- AlicenseNot gradedqualityCmaintenanceUnlimited web access MCP server for AI agents, enabling search, fetch, and learning across multiple sources while bypassing anti-bot protections.MIT
- AlicenseNot gradedqualityDmaintenanceMCP server for web scraping and browser automation, enabling AI agents to extract clean, token-efficient content from web pages.1MIT

aethyn-browser-mcpofficial
AlicenseAqualityCmaintenanceAn MCP server enabling AI agents to drive a real browser through residential proxies with per-task geo (country/city) and sticky identity, all chosen at call time.10633MIT
DLBrowserofficial
AlicenseNot gradedqualityBmaintenanceMCP server that gives AI agents reliable web access with self-healing, anti-bot bypass and clean content extraction. It provides 11 tools like fetch, scrape, and search, metered by credit with cost transparency.1MIT