TrustScore โ€” Confidential Credit Scoring and Salary Benchmarking on FHEVM [Season 2 Builder Track]

Hey Zama community :waving_hand:

I just submitted TrustScore to the Season 2 Builder Track and wanted to share it here for feedback.

What is TrustScore?

TrustScore is a confidential financial reputation protocol built on Zama FHEVM. It solves a real problem: 1.4 billion people across Africa and emerging markets have no formal credit history and are locked out of credit โ€” not because they are bad borrowers, but because sharing raw financial data creates surveillance and privacy risks.

TrustScore uses Fully Homomorphic Encryption to let anyone build a verifiable financial profile on-chain, privately and trustlessly.

Two modules

Module A โ€” Salary benchmarking
Workers submit encrypted salaries to a benchmark pool. The smart contract computes market averages across all submissions without anyone ever seeing individual numbers.

Module B โ€” Credit scoring
Borrowers submit encrypted income, savings, and repayment history. Lenders verify creditworthiness without ever seeing raw data.

How the privacy works

  1. Data is encrypted in the browser using the Zama Relayer SDK before it touches the blockchain
  2. The smart contract stores only euint64 ciphertext โ€” nobody can read the raw values
  3. The borrower grants a specific lender wallet address access via grantAccess()
  4. The lender signs an EIP712 message, the Zama relayer decrypts and returns only the authorized fields

Links

Tech stack

  • Solidity + FHEVM library (euint64, FHE.add, FHE.fromExternal, ACL)
  • Hardhat for development and testing (6 tests passing)
  • Next.js frontend with Zama Relayer SDK
  • Deployed on Ethereum Sepolia testnet

Would love any feedback from the community and the Zama team. Happy to answer questions about the implementation.

1 Like

Hey @quab, welcome to the community and thanks for submitting TrustScore to the Season 2 Builder Track! Weโ€™ve received it and the team will take a closer look soon.

Any questions in the meantime? Happy to help.