Zero-knowledge proofs, explained

Here is an everyday absurdity: to prove you are old enough to buy a drink, you hand over a card that shows your exact birth date, your full name, your address and your licence number. The question needed one bit of information — over 18, yes or no — and you disclosed a dossier. Every identity interaction built on documents has this shape: the evidence over-answers the question.

A zero-knowledge proof is the cryptographic fix. It lets one party (the prover) convince another (the verifier) that a statement is true — "this verified person is over 18" — while revealing nothing beyond the truth of the statement itself. Not the birth date. Not the document. Nothing but the yes.

Imagine a locked circular cave with two entrances, A and B, connected deep inside by a door only you can open. To prove you hold the key without showing it, you go in through one entrance while a verifier waits outside. They then shout which entrance you must come out of. If you cannot open the door, you have a 50% chance of already being on the right side by luck. Repeat twenty times and luck runs out: only someone with the key could keep emerging correctly. The verifier is now certain you hold the key — yet has learned nothing about the key itself.

Real zero-knowledge protocols replace the cave with mathematics, and the repeated challenges with equations that would be unsatisfiable unless the hidden value truly has the claimed property. The essential trick is the same: the proof demonstrates a consequence of the secret, never the secret.

Three properties make a proof "zero-knowledge" in the formal sense. Completeness: if the statement is true, an honest prover always convinces the verifier. Soundness: if it is false, no cheating prover can convince them except with negligible luck. Zero-knowledge: the verifier learns nothing they could not have computed on their own — the transcript of the proof is, informationally, worthless to anyone else.

The VEID layer that powers the Identity Wallet ships zero-knowledge circuits for the questions services most commonly over-collect on. Each circuit takes private inputs that never leave your control, checks them against a commitment anchored at verification time, and produces a proof the service can verify mathematically.

• Age range — proves your age is at or above a threshold (over 18, over 21) from your verified date of birth, without revealing the date itself.
• Residency — proves your verified address is in a given country, without revealing the address.
• Score range — proves your verification score exceeds a threshold a service requires, without revealing the exact score.

For you, proofs mean the everyday absurdity ends: an age-gated service gets a yes, and your birth date stays yours. For services, proofs are better evidence, not just kinder evidence — a cryptographic verification is harder to fake than a document photo and easier to check than a human judgement, and it leaves the service holding nothing breachable. The catalogue of what each proof demonstrates and what it never reveals is published as part of the service.

There is one honest caveat. A proof binds to the moment and the question; it does not stop a service asking for more than it needs through other channels. That is a consent problem rather than a cryptography problem, and it is why proofs sit inside a consent framework where every request states its purpose and every grant is revocable.

Zero-knowledge proofs invert an assumption baked into the information age: that verifying something requires seeing it. Once verification and disclosure are separated, whole categories of data hoarding lose their justification. Identity is the sharpest application because the data is the most dangerous to hoard — but the underlying principle, prove the property and keep the data, is one of the few genuinely new tools privacy has gained in decades. The wallet exists to make it ordinary.