ZK proof (Zero-Knowledge proof)

A zero-knowledge proof demonstrates that a statement satisfies a relation without revealing its secret witness. The verifier learns what is needed to check the statement under the chosen security model. Not every validity proof or application called ZK conceals transactions or their data.

Written by Pablo Gómez

Founder & CEO at Metlabs · Reviewed on October 11, 2026

A prover uses public inputs and a secret witness to produce a proof. A verifier checks the relation without receiving the witness. Zero knowledge limits disclosure through that proof; it does not hide information published elsewhere by the application.

SNARK and STARK families have different assumptions and costs. Some constructions require trusted parameters; others are transparent. PLONK with KZG commitments uses universal setup, while variants with other commitments have different requirements.

Proofs can verify attributes or computations. For scaling, a rollup may use validity proofs while publishing transactions, so privacy is not automatic. Construction and application require individual assessment.