Privacy-Preserving Academic Transcript Verification Using Zero-Knowledge Proof

Abstract

In admissions and hiring, transcript verification often requires only a threshold decision, for example a grade point average of 3.0 or higher or a Python grade of B or better. However, prevailing workflows demand full transcript disclosure, creating privacy risks and evaluation bias. We present a deployable system that proves a chosen subject meets a required threshold without revealing the score or any other subjects. The system integrates Groth16 Succinct Noninteractive Arguments of Knowledge (SNARKs) with W3C Decentralized Identifiers (DID) and Verifiable Credentials (VC) in a four-service architecture for DID registration, university issuance, student proving, and third-party verification. Holder identity is enforced offcircuit via a DID based challenge-response protocol where the student signs the verifier's nonce and the verifier checks that the signer's DID matches the VC's subject identifier. The zero-knowledge (ZK) circuit proves only Merkle inclusion and threshold comparison. The university signs the transcript's Merkle root with an Ed25519 signature. The verifier validates this signature using a key obtained from the university domain or a trusted registry, never from the student. On standard development hardware, steady-state proving latency is between 0.66 and 1.10 seconds. The verifier learns only a pass/fail bit, enforcing data minimization. In our negative test suite, no false accepts were observed.