proxy-scraper — Claude Code MCP server: what it does, install
Collects public HTTP, SOCKS4 and SOCKS5 proxies, checks them, and hands the agent the outcome of each check instead of a single 'working' label.
claude mcp add proxy-scraper -- uvx proxy-scraper-cliWhat proxy-scraper does
It gives the agent the same thing the command line does: proxy lists and the result of their checks. Addresses come from public lists, each one goes through up to four checks, and every check keeps its own outcome: passed, failed, unavailable or skipped, with a reason code, the target and the time. A check that never ran is not reported as a pass.
Picking a proxy is a short loop. --pick pulls some from the published list, --verify retests them from your own network, --diagnose walks one socks5://HOST:PORT through every check. Exports exist for curl, proxychains, Clash and sing-box; there is also a local rotating proxy and a Python API.
You need uv on the machine, since uvx ships with it. Both --verify and --diagnose arrived in version 1.30. Free proxies are run by strangers, and only about a quarter of an hour's list shows up again a day later, so treat one as fine for public pages and test runs, not for logins or payments.
What it helps with
- grab a proxy that answers from your network right now
- explain why one SOCKS5 proxy keeps failing
- recheck your own proxy list before a scrape run
- export working addresses for curl or proxychains
- spin up a local rotating proxy for tests
Example prompts for the agent
- Pull three HTTP proxies from the fresh list and verify them from my network.
- Work out why
socks5://HOST:PORTfails and show me each check it did and did not pass. - Recheck the proxies in my file and keep only the ones that answer over HTTPS.
How to install proxy-scraper
Platform CLI
claude mcp add proxy-scraper -- uvx proxy-scraper-cliuvx runs the command named after the package; if the README names another one, use uvx --from proxy-scraper-cli <command>
Platform CLI
codex mcp add proxy-scraper -- uvx proxy-scraper-cliuvx runs the command named after the package; if the README names another one, use uvx --from proxy-scraper-cli <command>
Config file
{
"mcpServers": {
"proxy-scraper": {
"command": "uvx",
"args": [
"proxy-scraper-cli"
]
}
}
}uvx runs the command named after the package; if the README names another one, use uvx --from proxy-scraper-cli <command>
Config file
{
"servers": {
"proxy-scraper": {
"type": "stdio",
"command": "uvx",
"args": [
"proxy-scraper-cli"
]
}
}
}uvx runs the command named after the package; if the README names another one, use uvx --from proxy-scraper-cli <command>
Platform CLI
gemini mcp add proxy-scraper uvx proxy-scraper-cliuvx runs the command named after the package; if the README names another one, use uvx --from proxy-scraper-cli <command>
Step by step, with commands for each platform: How to add an MCP server to Claude Code, Codex, Cursor, VS Code, Gemini CLI and more
Questions about proxy-scraper
How do I install proxy-scraper in Claude Code?
With one command: claude mcp add proxy-scraper -- uvx proxy-scraper-cli. Have uv installed first, because uvx comes with it.
What does proxy-scraper do?
It gathers public HTTP, SOCKS4 and SOCKS5 proxies and passes the agent a list with the result of each check. A result is passed, failed, unavailable or skipped.
What do I need to run proxy-scraper?
uv has to be there, since uvx is part of it. --verify and --diagnose need version 1.30 or newer.
