Cred Protocol

Cred Protocol is an on-chain credit risk quantification protocol whose Andromeda machine-learning model converts wallets' borrowing history, repayment behavior, asset composition, protocol diversity and identity attestations into a Cred Score from 300 to 1000, with addresses lacking borrowing history still scored on their on-chain behavior, covering more than 300 million EVM addresses, 10 chains and over 30 lending protocols including Aave, Compound, Morpho and Spark. It targets the inability of overcollateralized lending to identify quality borrowers: scores let protocols lower collateral requirements for low-risk addresses, while credit reports and real-time webhooks push liquidation, default and delinquency events to integrators, delivered through APIs and SDKs.

In on-chain reputation and identity, Cred Protocol complements networks such as Quadrata, Orange Protocol and Krebit, with its identity endpoint aggregating ENS, Gitcoin Passport, POAP and Worldcoin credentials and sybil detection analyzing funding-source clustering and transaction timing. For the AI agent economy it offers 21 MCP tools for on-chain reputation analysis. Model features are not fully disclosed to prevent targeted score gaming.