Install one skill. Your AI agent gets proxy access to public websites through real residential IPs — with country routing and sticky sessions. Advanced integrations require separate setup and verification.
AI agents need web data. But the web doesn't want them. CAPTCHAs, bot detection, IP bans, JavaScript walls, geo-restrictions — every website is designed to block automated access. Your agent sees error pages where data should be.
Full web access through the world's most advanced residential proxy infrastructure.
Rendering is a separate integration. The V2 proxy quickstart returns target responses without executing JavaScript.
CAPTCHA handling is a separate integration, not guaranteed by the V2 proxy connection.
Residential proxy routing for public web requests. Target acceptance varies; no universal bypass is guaranteed.
Access content from any location. Country targeting is available. State/city targeting is deployed in V2; exact-city accuracy varies.
Get clean, structured output ready for your LLMs. HTML to markdown conversion built in. Structured JSON for search results.
Proxy requests consume bandwidth. Contribution programs and credits are separate; check terms with support.
Proxy traffic consumes account bandwidth. Ask support about contribution eligibility and terms before participating.
Contribution programs have separate eligibility and terms; fetching does not promise credits
Confirm node software, consent and program terms before sharing bandwidth
Requests consume bandwidth; check your dashboard balance
ProxyClaw routes your agent's requests through real residential IPs — real devices, real browsers, real fingerprints. Target acceptance varies; inspect each response.
One command. Your AI agent gets instant access to IPLoop's residential proxy network — residential routing; availability varies by country.
Use V2 for public-page requests and country routing. Rendering and extraction require separate integrations.
Contribution programs and credits require separate eligibility. Check account balance and pricing before use.
ProxyClaw ships as a Model Context Protocol server. Give Claude Desktop, Cursor, or Zed a real residential IP across Live country availability — no scraping code. Target acceptance varies.
Clone and build the updated MCP source for V2 credentials.
Setup on GitHub →Install and start fetching in under 30 seconds.
Target acceptance varies. V2 HTTP, SOCKS5 and sticky examples are in the docs. GitHub →
| Parameter | Values | Default | Description |
|---|---|---|---|
url | any URL | (required) | Target page to fetch |
format | rendered, raw, markdown | rendered | Output format |
country | 2-letter code | auto | Geo-target country |
expiration | 0 to N (days) | 1 | Cache age (0 bypasses) |
delay | 0.1 to 10 (sec) | none | Extra wait for dynamic content |
device | desktop, mobile | desktop | Device type |
ip | residential, isp | residential | IP type |
View current pricing, balance and plans in your dashboard. No free quota or contribution credits are promised by these examples.
Open DashboardV2 proxy access with explicit credentials and account bandwidth. Install ProxyClaw and start scraping in 30 seconds.