The one-sentence version
A zero-knowledge proof lets you prove a statement is true without revealing why it is true. That is the entire idea. You convince a verifier that a claim holds — “this transaction is valid” — while the evidence behind the claim stays with you. The verifier learns exactly one thing: true. Nothing else leaks — not the UTXOs behind the claim, not your balance, not your history, not the reason the claim holds.Two ways to picture it
Over 18, no birthdate
Buying a drink today means handing over an ID card that shows your full birthdate, your address, and a document number — far more than the question required. A zero-knowledge version would confirm one fact: over 18, yes. The bartender learns the single thing they needed and nothing they didn’t. Cloak applies the same principle to money: the chain learns “this spend is legitimate” — not which UTXO funded it, what it held, or which wallet deposited it. The link back to you is what stays sealed.Sealed envelopes in a public mailbox
Shielded pools are like sealed envelopes in a public mailbox: everyone sees mail, not what’s inside. A zero-knowledge proof is what lets you take value back out without opening anything in public. You prove that one of the envelopes is yours and still sealed — without pointing at which one.What the proof actually proves in Cloak
When you shield funds, they become UTXOs — the individual private notes inside your private balance. Every time you spend from that balance, your browser produces a proof that convinces the Cloak program of three things:- The UTXO you’re spending is really yours. Only someone holding its secret can produce the proof, and the proof never reveals which UTXO in the pool it refers to.
- It hasn’t been spent before. Each spend publishes a one-time marker called a nullifier. Spend the same UTXO twice and the marker repeats, so the program rejects the transaction — still without learning which UTXO was involved.
- The amounts balance. What leaves your UTXOs equals what arrives on the other side — a payment and, usually, change back to you. Nothing created, nothing destroyed. A transaction can never invent value that was never deposited.
Where it happens
In your browser. When you confirm a private send, your device generates the proof locally, typically in under 3 seconds, and attaches it to the transaction. The chain verifies the proof, never the details. Your UTXO secrets are never uploaded, never sent to a server, and never seen by Cloak.Why this isn’t “trust us”
The Cloak program accepts nothing without a valid proof — not from you, not from Cloak’s own infrastructure, not from anyone. There is no override, no allowlist, no path where a transaction goes through because the right party asked nicely. The rules are enforced by verification, not reputation. The program itself is upgradeable — how that is governed, and everything else Cloak cannot do, lives on the Security page. That cuts both ways, and honestly so. Cloak can’t see your balance, can’t spend it, and can’t recover it. That’s the design. The same proofs that stop a stranger from spending your UTXOs also stop us. One breadcrumb for the curious: Cloak’s proofs are Groth16 proofs, and the pool’s data structures use Poseidon hashing. The full construction lives in the shield pool reference.Where next
How Cloak works
The four flows — shield, send, swap, unshield — and what happens in each.
Your private balance
UTXOs, the backup file, and what happens if you lose them.