Loops HouseBlindroll

NEAR HACKATHON
Links
Team
1 member- TUOwner
Tushar Singh Chouhan
Overview
Problem: paying a team in stablecoins on a public chain creates a public salary table, and real payroll sheets are messy (different column names, number formats, wrong addresses).
Blindroll is claim-based USDC payroll on NEAR. An approver uploads the payroll CSV straight from an HR or accounting export. A TEE-hosted model on NEAR AI Cloud proposes the column mapping; the response is signed and checked, and deterministic rules flag errors and outliers (the AI never decides a payment). The cleaned table becomes a Merkle tree: only the Merkle root and an attested report hash go on-chain, so there is no public salary table. Approvers sign on-chain (M-of-N) and fund the run with USDC. Each contributor opens one link, proves their wallet (NEAR or MetaMask), sees only their own row, and claims gas-free through a relayer; the NEAR contract verifies Merkle proofs and EVM signatures. NEAR Intents (1Click) quotes show cross-chain payout options (delivery simulated on testnet). Judges can try it end to end in a private 1-of-1 sandbox on the hosted testnet app.