Proof of humanity with zero-knowledge proofs
- Problem
- Bots drained airdrops and DAO votes; KYC was not an option.
- Built
- A feasibility assessment, then a proof-of-concept prototype.
- Scale
- Six trust levels, L0 to L5
A zero-knowledge proof lets a user or a system prove that a statement is true (a score is above a threshold, a person holds a valid credential, a batch of transfers is correct) without revealing the data behind it. We design the circuits, the proving flow on the client or a relayer, and the smart contracts that verify the proofs, and we built and maintained the zkSync SDKs in five languages.
| Proof systems | Groth16 · Halo2 · recursive proofs |
|---|---|
| Verification | Ethereum smart contracts |
| Layer 2 | zkSync SDKs: Go, Python, Rust, Android, iOS |
| Used for | Sybil resistance · private auth · scaling |
How a project runs
| Threshold proofs | A user proves "my score is at or above the threshold for this level" without revealing the score or any input, verified in a smart contract. Feasibility assessment and a proof-of-concept prototype for an anonymous proof-of-humanity protocol with six trust levels. | |
|---|---|---|
| Private authentication | Zero-knowledge authentication and data checks for a private messenger with seed-phrase identity, end-to-end encryption and encrypted metadata. | |
| ZK rollup integrations | The full set of zkSync SDK libraries for integrators, built and maintained by our team in Go, Python, Rust, Android and iOS, later with EIP-712 signatures and deterministic deployments. | |
| Off-chain to on-chain flows | Commitments from an off-chain engine (for example an ML risk score), proofs generated on the client or by a relayer, and on-chain verifiers that partner contracts can call. | |
| Proof system choice | Groth16 where cheap verification on Ethereum matters most, with a planned path to Halo2 and recursive proofs where a trusted setup or proof aggregation is the constraint. | |
| Privacy without KYC | Designs that replace passports and biometrics with device attestation, behaviour signals and proofs, so a protocol never stores who the user is. |
Stack
Proof
FAQ
When a third party must trust a fact but must not see the data behind it: a trust level without the score, eligibility without identity, a correct batch without replaying it. If the data can be shown, a signature or a plain on-chain record is simpler and cheaper.
Groth16 gives small proofs and cheap verification on Ethereum, but needs a trusted setup per circuit. Halo2 avoids that setup and supports recursion, at the cost of larger proofs. Many projects start on Groth16 and plan the move once circuits settle.
Yes, for small circuits. For larger ones the proof is generated by a relayer, and the design must make sure the relayer learns nothing it should not.
Yes. Our team built and maintained the zkSync SDKs in five languages and advised zkSync integrators on their integrations.
Next step
A 30 minute call with Mykhailo. You leave with a clear view of scope, risks and next step.
Mykhailo Tiutin
CTO, Vareger
