Loops HouseIN
InferencBid

Event
NEAR HACKATHON
Links
Team
1 member- KEOwner
Kelsen Liu
Overview
Problem: Agents need to purchase inference within budgets, balance price against latency, and verify the model service they receive. Connecting this demand with available capacity is an opportunity for NEAR AI.
Solution: We built InferenceBid, a private inference marketplace with remote TEE attestation supporting our goal of provable inference and model fidelity. Our proposal adds autonomous bidding and on-chain payments, potentially increasing Cloud revenue and capacity utilization while enabling Gateway agents to procure inference independently. We use NEAR AI Cloud, Chutes, Phala CVMs, and OutLayer wallets, with MPP and NEAR Intents proposed for payment and funding flows.