CAPTCHA

CAPTCHA stands for Completely Automated Public Turing test to tell Computers and Humans Apart. The term covers any challenge – distorted text, image grids, behavioral analysis, or invisible cryptographic puzzles – that a website presents to gate access against automated traffic. Google's reCAPTCHA, hCaptcha, Cloudflare Turnstile, and FunCaptcha are the most widely deployed implementations. Modern CAPTCHAs are usually invisible to legitimate users and only escalate to a visible challenge when the underlying behavioral risk score crosses a threshold.

For web scrapers, CAPTCHAs are one of the highest-friction blockers. The two practical responses are avoidance and solving. Avoidance means reducing your bot signature so the CAPTCHA never triggers – better proxies, stealth browser configurations, lower request rates, more realistic interaction patterns. Solving means handing the CAPTCHA to a human-powered or ML-powered service (2Captcha, Anti-Captcha, CapSolver) that returns the answer for a fee, typically $1–3 per thousand solves for image CAPTCHAs and $0.50–2 per solve for reCAPTCHA v2/v3.

For AI builders, the strategic choice is to use a scraping platform that handles CAPTCHA encounters automatically – most do – rather than building your own solving pipeline. Services like ZenRows and Scrapfly bundle CAPTCHA solving into the per-request price; others charge à la carte. Either way, treating CAPTCHA encounters as a cost-of-business and routing them through infrastructure is far less work than fighting them yourself.