Summary of an article published on The Web Scraping Club on 2025-05-25 (free to read). Read the full article

Paid proxies cost money, and free online proxy services can’t be trusted. A few years ago I wondered whether Tor, which I knew only as a way to browse anonymously, could anonymize my IP while scraping. It can. Run as a local service, it behaves like a free rotating proxy that changes exit IP every few seconds, which helps against rate limiters and IP bans.

Here I show the setup I followed. It uses the Tor daemon as a system service, not the Tor Browser, and exposes a local SOCKS proxy. We check that it works against HTTPBin’s /ip endpoint, first with cURL, then from Python with Requests and from JavaScript with Axios. I also tell the story of a client project with millions of rate-limited pages, where I used Tor instead of paid proxies, and I list why Tor is a poor fit for production pipelines.

What’s inside

  • The Tor daemon running as a system service is a different thing from the Tor Browser, and the setup differs by operating system.
  • A SOCKS proxy works at a lower level than an HTTP/HTTPS proxy and can carry other protocols too.
  • The torrc file lets you change how often Tor builds a new circuit, with a trade-off in connection stability and load on the Tor network.
  • Limits compared with premium providers: little control over rotation, no choice of exit country, slower and more error-prone connections, and publicly listed exit IPs.

In the full article: The Python and JavaScript code, the cURL checks per operating system, the torrc setting, and the full story of the multi-instance Tor setup I used on a client project.

Read the full article on The Web Scraping Club