zkHealth

MedZK is a privacy-first medical insurance claim platform that lets insurers verify eligibility without seeing a patient’s raw medical data. Instead of sharing full lab reports, an authorized medical lab generates a zero-knowledge claim bundle that proves the patient satisfies specific medical conditions, such as high blood sugar or high cholesterol, while keeping the exact values private. The insurer only learns that the criteria were met, not the underlying medical record.
In the current workflow, the lab uploads a medical report, generates the proof locally, and signs the claim bundle with its wallet. That signature proves the bundle came from an authentic issuer. The signed bundle can then be shared with the insurer through a file, QR code, or direct link. On the insurer side, the bundle is loaded into a verification portal where the insurer can confirm the proof, validate that the bundle was signed by a registered lab, and review claim-safe public metadata without accessing the patient’s sensitive health information.
MedZK also includes an optional on-chain workflow on Sepolia. Signed claim bundles can be submitted to a smart contract, and Chainlink Automation can monitor and process pending claims automatically. This creates an auditable and automatable path for claim handling while preserving privacy. MedZK shows how zero-knowledge proofs, authenticated medical issuers, and Chainlink-based automation can work together to reduce over-disclosure, improve trust, and modernize medical insurance claims.

MedZK is built with Next.js, Tailwind CSS, Framer Motion, and a Noir-compatible zero-knowledge proof flow running locally in the browser. Labs upload medical reports, generate claim proofs, and sign claim bundles with MetaMask so insurers can verify both eligibility and issuer authenticity without seeing raw patient data. Signed bundles can be shared by file, QR code, or link, and optionally submitted on-chain to a Solidity ClaimManager contract on Sepolia, where Chainlink Automation can process claims automatically.