Lookout is an onchain reputation scoring layer for AI agents. It audits agent behavior from public onchain transactions, integrates ZK identity verification via Self Protocol, and writes composable TrustScores that any agent, dApp, or human can query before transacting.

Lookout is the credit score for AI agents. It builds on top of ERC-8004 to fill the gap between agent identity and agent trust — providing automated, objective, onchain reputation scoring that any agent, dApp, or human can query before paying, trusting, or cooperating. Deployed on Celo and Base, powered by Self Protocol ZK identity and x402 payments. One question answered: should your agent trust this other agent?

AI agents are entering the onchain economy, transacting, paying, and cooperating autonomously. But when your agent needs to interact with another agent, there's no way to verify if it's trustworthy. ERC-8004 provides identity and a space for reputation, but intentionally leaves scoring strategies external. The result: agents can register, claim to be reliable, and start transacting, with zero accountability. There's no automated, objective way to assess an agent's behavior before trusting it with your money.

Lookout is the missing scoring layer. It autonomously audits any agent's onchain behavior — transaction history, success rates, counterparty diversity, account age, activity patterns — and writes a composable TrustScore (0-100) onchain. Agents verified as human-backed via Self Protocol ZK proofs get a +15 bonus. Any agent can query a TrustScore via a single API call before deciding to transact — no API key, no registration, no human in the loop. Audits are paid via x402 micropayments ($0.01 USDC), making Lookout a self-sustaining autonomous service.

Pay-per-audit via x402 micropayments. Agents pay $0.01 USDC per audit request. Revenue covers onchain gas costs, making Lookout a self-sustaining autonomous service. Score reads are free, audits are paid.

Want to grow through just grant funding

Global