I’ve been writing my own Playwright scripts for a while, but recently I’m getting instantly blocked by Cloudflare and Datadome on almost every target site. I’ve tried rotating proxies, but it seems they are detecting the browser itself.
I read in the docs that RTILA uses something called patchright. How is this different from standard Playwright?
Standard Playwright is fantastic for testing, but it leaks browser fingerprints everywhere (like navigator.webdriver = true), making it trivial for Datadome or Cloudflare to flag you as a bot.
Instead of standard Playwright, our Deno engine uses patchright. This is a heavily modified version of the browser binaries designed specifically to strip out those automation flags at the source code level.
But we don’t stop there. Even with a clean fingerprint, bot systems track behavior. That’s why our engine includes a dedicated humanoid interaction service. Instead of instantly teleporting the cursor to a button, it calculates complex bezier-curve mouse trajectories, adds randomized typing delays, and simulates human hesitation. Combine that with your proxy rotation, and you look exactly like a real user.
Standard Playwright is fantastic for testing, but it leaks browser fingerprints everywhere (like navigator.webdriver = true), making it trivial for Datadome or Cloudflare to flag you as a bot.
Instead of standard Playwright, our Deno engine uses patchright. This is a heavily modified version of the browser binaries designed specifically to strip out those automation flags at the source code level.
But we don’t stop there. Even with a clean fingerprint, bot systems track behavior. That’s why our engine includes a dedicated humanoid interaction service. Instead of instantly teleporting the cursor to a button, it calculates complex bezier-curve mouse trajectories, adds randomized typing delays, and simulates human hesitation. Combine that with your proxy rotation, and you look exactly like a real user.