Loops HouseAgent-Receipt

NEAR HACKATHON
Links
Team
1 member- RUOwner
Rushikesh Ghule
Overview
I built AgentReceipts, an evidence-first financial control and audit platform for AI agents on NEAR.
It allows an external AI agent to request payments while keeping humans in control. Before a payment can happen, AgentReceipts checks spending policies such as per-transaction limits and daily budgets. The user then reviews and explicitly approves the payment through a connected NEAR wallet.
After approval, the transaction is executed on NEAR Testnet and verified against on-chain data.AgentReceipts generates a human-readable receipt containing the payment amount, sender, recipient, transaction hash, network fee, timestamp, and verification status.
The goal is to make AI agent spending controlled, transparent, and independently verifiable.