Working with reCAPTCHA in automation
How reCAPTCHA behaves when the visitor is a script, what a v3 score actually means, and how to handle it in an agent or test suite you are authorized to run.
reCAPTCHA v3 never challenges your agent. It scores it, and the site's own threshold rejects the request. How to diagnose that, and what to do next.
Read →Detect the challenge, POST /solve for a token, inject it where the page expects it. Working Playwright and Puppeteer examples with real request shapes.
Read →v2 returns a pass/fail token. v3 returns a score the site grades itself. Enterprise changes the verification path. What each means for your integration.
Read →For reCAPTCHA, only if the task name doesn't end in ProxyLess — the other task families have their own rules. The exact check the KagedCap API runs, the proxy formats it accepts, and why the mint IP changes your v3 score.
Read →