> ## Documentation Index
> Fetch the complete documentation index at: https://mr.verifyyou.com/llms.txt
> Use this file to discover all available pages before exploring further.
# What HumanCheck is
> HumanCheck is VerifyYou's camera-based check that confirms a real, unique, live person is behind a response, without collecting a name or an ID. Here is how it works, for market research.
HumanCheck is VerifyYou's camera-based way to tell a real, unique respondent from a bot, a duplicate, or an AI. It confirms two things: that a real, live person is present right now, and that this person is unique, not someone taking your study more than once under different accounts. It does all of that without collecting a name, an address, or a government ID.
## The one question it answers
HumanCheck answers a single question: is there a real, unique, live human on the other end of this response, right now? It does not try to learn who someone is. It focuses only on confirming that a genuine, individual person is present. A respondent points their camera at their face for a brief scan, a few seconds in the browser, with no documents and no government ID.
## The two things it proves
A real person is present in the moment, which is what makes bots and fake media hard to get through. The check looks for the small signals that only a living person produces right then, and is built to reject photos, recordings, a screen showing a screen, and AI generated faces.
One human, one response. This keeps a single person from quietly taking the same study many times under different accounts. When a genuine respondent returns, the check can recognize them, while someone you have removed for cause finds it hard to come back under a new identity.
## The return journey
Verification is not only a one-time gate. It works across visits, so you can be strict where it matters and light where it does not.
The initial check takes seconds and produces a verdict from the face scan: verified or denied.
On later visits you choose how strict to be. Require a fresh scan every time, or quietly recognize a returning respondent by device, so a genuine repeat participant is not made to prove themselves again.
A respondent who confirmed an email or phone during the check can be recognized across devices. Guests, who did not confirm either, are recognized more lightly, and their records fade over time.
## What it is not
It does not verify a legal identity or a name, on purpose. For survey work you almost never need to know who someone legally is, only that they are a real person you have not already counted.
Two-factor proves that someone holds a credential, like a phone or an email. HumanCheck ties the check to the actual person in front of the camera, not to a thing they possess.
It asks for a live presence, which is a stronger signal than passive behavioral analysis.
## Your respondents' data
The scan becomes a face vector, a one-way numerical representation of the face's geometry, used only to recognize a returning person. It is not a stored image. Your panel and your platform never touch any of that. What you receive is a verdict, verified or denied, and the reference id you chose to attach, never a face and never a name.
## Where it fits in a study
Put the check right before the incentive. Real people push through because they want to finish, and the accounts that balk are usually the ones you were about to pay for nothing. It sits alongside the checks you already run, your attention checks and your device signals, and answers the real-versus-fake question those were not built to settle.
Liveness and uniqueness in a few seconds, and the honest boundary.
The real screens a respondent sees, from click to complete, and the demo.