zk-STARK
What is zk-STARK?
zk-STARKA Zero-Knowledge Scalable Transparent Argument of Knowledge: a post-quantum-friendly zero-knowledge proof system that needs no trusted setup and relies only on collision-resistant hash functions.
A zk-STARK (Zero-Knowledge Scalable Transparent Argument of Knowledge) is a non-interactive proof system that lets a prover convince a verifier a computation was executed correctly, while revealing nothing beyond its validity. It was introduced in 2018 by Eli Ben-Sasson, Iddo Bentov, Yinon Horesh and Michael Riabzev (IACR ePrint 2018/046, "Scalable, transparent, and post-quantum secure computational integrity"), and commercialised by StarkWare.
A computation is first expressed as an Algebraic Intermediate Representation (AIR) — a set of polynomial constraints over an execution trace. The trace is low-degree extended, committed to via Merkle trees, and proven close to a valid low-degree polynomial using FRI (Fast Reed–Solomon Interactive Oracle Proof of Proximity). The Fiat–Shamir heuristic makes the protocol non-interactive.
Unlike zk-SNARKs, STARKs need no trusted setup ("transparent"), eliminating the toxic-waste risk of a setup ceremony, and their security rests only on collision-resistant hashes — making them plausibly post-quantum secure. Prover time scales quasi-linearly and verification is polylogarithmic. The trade-off is larger proofs, typically tens to hundreds of kilobytes. Recursive STARKs (proofs of proofs) compress many transactions into one, and zk-STARKs underpin StarkEx, Starknet and the Cairo language, plus verifiable off-chain and ML-inference workloads.
flowchart LR A["Computation"] --> B["AIR: polynomial constraints<br/>over execution trace"] B --> C["Low-degree extension<br/>+ Merkle commitment"] C --> D["FRI low-degree test"] D --> E["Fiat-Shamir:<br/>non-interactive proof"] E --> F["Verifier checks<br/>in polylog time"]
● Examples
- 01
Starknet uses zk-STARKs to prove the correctness of off-chain transaction batches against Ethereum.
- 02
Verifiable ML services that prove a model produced a specific output without revealing the inputs.
● Frequently asked questions
What is zk-STARK?
A Zero-Knowledge Scalable Transparent Argument of Knowledge: a post-quantum-friendly zero-knowledge proof system that needs no trusted setup and relies only on collision-resistant hash functions. It belongs to the Cryptography category of cybersecurity.
What does zk-STARK mean?
A Zero-Knowledge Scalable Transparent Argument of Knowledge: a post-quantum-friendly zero-knowledge proof system that needs no trusted setup and relies only on collision-resistant hash functions.
How do you defend against zk-STARK?
Defences for zk-STARK typically combine technical controls and operational practices, as detailed in the full definition above.
What are other names for zk-STARK?
Common alternative names include: STARK, Scalable Transparent Argument of Knowledge.