Zegent

Zegent enables borrowers to access undercollateralized loans using zero-knowledge proofs. Traditionally, high collateral requirements restrict borrowing, while lenders lack visibility into borrowers’ financial status. By leveraging Circom and zk-SNARKs (Groth-16), Zegent allows borrowers to prove their financial eligibility without revealing sensitive data, enabling lenders to offer more flexible loan terms.

Beyond lending, Zegent enhances token trading and analysis across multiple exchanges and protocols through its verifiable AI agent, CryptoSage, built on the CDP AgentKit. To strengthen its capabilities, we have developed additional tools alongside existing ones for seamless trading and analysis. CryptoSage is secured using EigenLayer’s Autonomous Verifiable Service (AVS), ensuring trust and transparency. When users submit queries, the AgentServiceManager contract triggers a NewTaskCreated event, notifying registered Operators. These Operators, who have staked and delegated assets, process the task independently. Once completed, they sign and submit a hashed response using their private key back to the AVS contract. This decentralized mechanism guarantees secure, verifiable AI-powered responses, enhancing both trust and efficiency in crypto transactions and loan processing within the Zegent ecosystem.

How Undercollateralized Lending Works

Zegent enables undercollateralized lending by assessing borrowers' financial history from various exchanges and protocols using the Arbitrum API for subgraphs deployed on GraphQL. The retrieved financial data is processed through the generate_witness algorithm in zk-SNARKs, which generates private and public signals based on parameters set in the Circom circuit. Borrower history remains private, while loan threshold eligibility and other constants are public.

The public signals and zero-knowledge proof are then sent to the Loan Contract, which inherits the Verifier Contract to call the verify function. This function returns a boolean value indicating whether the borrower meets the loan eligibility criteria. Borrowers who pass the verification can create loan requests, and lenders can view only those who qualify based on the threshold score. Borrowers with weaker financial scores, who do not qualify for loans, have the option to use Zegent’s AI-powered exchange tools, leveraging exchange ABIs to strengthen their profile through analysis and informed trading.

CryptoSage – Verifiable AI for Trading & Analysis

CryptoSage, built using the CDP AgentKit, enhances token trading and analysis through AI-driven tools such as AaveQueryTool, AaveSupplyTokenTool, MorphoQueryTool, and AerodromeTool. To ensure verifiability, CryptoSage is secured through EigenLayer’s Autonomous Verifiable Service (AVS) mechanism.

When an AVS consumer queries CryptoSage, the request is sent to the AgentServiceManager contract, which emits a NewTaskCreated event. This notifies registered Operators, who have staked and delegated assets, to process the request. Each Operator independently hashes the response, signs it using their private key, and submits the signed hash back to the AgentServiceManager AVS contract. Only registered Operators who meet the required stake threshold can submit responses, ensuring a secure, verifiable, and decentralized AI-powered system for lending decisions and crypto analysis.