AgentFeed Earns a 77.5 Proof of Usefulness Score With Machine-to-Machine Crypto Data Payments

AgentFeed has achieved a Proof of Usefulness score of 77.5 in the recent Hackathon. Developed by Sergiu, it is a live crypto market data API that enables AI agents to pay for data via micro-transactions using the x402 protocol. AgentFeed eliminates the need for human intervention or legacy subscriptions by recording and billing only actual liquidation events from 799 perpetual futures on Bybit, OKX, and Binance.

With 50 tools available, including five free, AgentFeed runs on Node.js and SQLite on a single VPS, providing a simple, honest database of real-world data. The API's 77.5 score reflects its technical novelty, market timing, and completeness, despite the developer's near-zero audience engagement. AgentFeed's primary use is for developers building liquidation-aware trading agents, offering them a reliable, programmable, and independently paid-for data source.