Proof of Human's reported detection lead versus reCAPTCHA rests on behavioral biometric scoring—capturing cognitive signatures that are expensive for attackers to imitate at scale—rather than visual challenges that AI models can defeat. Its testing found 87% bot detection versus reCAPTCHA's 69%, with no user friction. In September 2024, researchers at ETH Zurich announced an AI model which defeated visual reCAPTCHAs 100% of the time.
Proof of Human does not collect PII and does not perform cross-site tracking. CNIL, the French data protection authority, determined that reCAPTCHA uses excessive personal data for purposes other than security.
AI agents that complete surveys and tasks undetected could lift Proof of Human's growth because the research organizations it serves must confirm that respondents are human; such agents can browse sites, finish tasks, and produce answers increasingly hard to separate from real people's, and they can now enter survey panels unnoticed, as the withdrawn YouGov study showed. Customer deployments confirm that demand: Stanford Psychology uses Proof of Human to verify that online study participants are real, and GrowProgress uses it for quality checks in polling and public-opinion research.
In a five-question survey, Proof of Human identified 99% of automated sessions, against 69.4% for reCAPTCHA v3, 44.3% for hCaptcha, 42.1% for Cloudflare Turnstile, and 9.1% for Cloudflare Bot Management. Its Community plan is free and offers complete visibility into bot checks in aggregate. Paid plans add session-level detail that lets customers inspect individual sessions in the dashboard and retrieve their data through the API.
Unmeasured false positives could impair the 99% detection rate that Proof of Human reported for its bot detector: the benchmark included no human sessions and measured detection of automated traffic without evaluating false-positive rates among human respondents, so a rising false-positive rate would signal that gap.
The benchmark ran automated browsers that loaded the anti-bot JavaScript, exposed their environment, sent telemetry, and waited to be detected. It measured publicly accessible configurations against a small number of isolated automated sessions on a new website with almost no historical traffic. Proof of Human evaluated the product tiers available to it rather than enterprise versions that may incorporate additional signals or more advanced detection.
Proof of Human's API lets website and research teams check that users are human invisibly in the background as they use a site normally, with no CAPTCHA, puzzle, or interruption, instead of a separate challenge, and it detects bots. It tracks behavioral biometrics—keystroke dynamics, mouse movements, and scrolling patterns—continuously, unlike traditional methods that depend on challenges or external data sources.
Proof of Human integrates behavioral, device, and network signals into a real-time assessment that companies apply when deciding whether to trust a given account, accept a given response, or pay for work. Customers place the Proof of Human JavaScript tracker on their site with one line of code. The Proof of Human Community plan is free and offers full observability into aggregate bot checks.