Summary of an article published on The Web Scraping Club on 2025-08-17 (free to read). Read the full article

Most tools used to build scraping bots produce generic, predictable fingerprints. Playwright, Puppeteer and Selenium spin up browser environments that are almost identical from bot to bot, so the fingerprint they send is extremely common. Real users personalize their browsers, and that is what anti-bot systems use to tell a human from a script. Cloudflare is the example: it collects data from the browser, analyzes it server side, and serves Turnstile when the fingerprint looks shady.

I explain what goes into a fingerprint, from JavaScript APIs to HTTP headers, and why fingerprints are also used for tracking across sessions. Then we work in Playwright. We generate the fingerprint of a vanilla instance on Am I Unique?, change the browser context settings, and test it again. I also cover the spoofing libraries, and why I now prefer scraping-specific automation tools like Camoufox, SeleniumBase and NoDriver, plus commercial anti-detect browsers like Kameleo and GoLogin.

What’s inside

  • Why the red bars on Am I Unique? do not mean you are safe: rare values stand out as suspicious, because the data comes from real human users.
  • Why you should run the fingerprint test on the server where the scraper will be deployed, not on your own computer.
  • Why a fixed fingerprint is a risk even when it looks convincing: User Behavior Analytics can link sessions, so rotation and proxy rotation go together.
  • Why a single bot-like data point, such as the user agent, can get you blocked however unique the rest of the fingerprint is.

In the full article: The full Playwright scripts for the vanilla and customized instances, the context settings, the Am I Unique? screenshots before and after, and the list of fingerprint spoofing libraries.

Read the full article on The Web Scraping Club