CAPTCHA

How it works

The name stands for “Completely Automated Public Turing test to tell Computers and Humans Apart”, coined by Luis von Ahn and colleagues in 2003. Early captchas were puzzles: distorted letters, then image grids. As software got better at puzzles, the services moved most of the work out of sight. Today’s widely used captchas, such as Google reCAPTCHA, hCaptcha and Cloudflare Turnstile, first look at the browser, the network and the way the visitor behaves, and only show a puzzle, or block, when that check is unconvincing.

So a captcha is usually the visible end of a risk decision rather than a fixed gate. The same page can wave one visitor through and challenge the next.

Why it matters when you run several accounts

Frequent captchas tell you a profile looks risky to the sites it visits. The usual reasons are the network (an address with a poor reputation, on a blocklist or in a hosting range), a browser that contradicts itself (time zone, language and IP in different places), signs of automation, or behaviour that is faster and more regular than a person’s. Captchas exist to stop abuse; a profile meeting many of them is worth fixing, not forcing through.

How SoftGlaze Browser handles it

SoftGlaze Browser doesn’t promise fewer captchas. It gives you the checks that find the usual causes. Each profile has a CAPTCHA risk score out of 100 that lists the reasons, and the profile list can be sorted and filtered by it; safe fixes, such as matching the time zone, language and location to the proxy or choosing a safer WebRTC mode, apply to one profile or many. Each proxy check looks the exit address up on 3 DNS blocklists and grades it, and Leak Check scores a profile’s consistency. From version 1.0.22, SoftGlaze AI (Beta) adds a prompt, “Why does this profile get CAPTCHAs?”, that reports what to fix. The acceptable use policy sets out what is and isn’t acceptable around challenges.