Trust infrastructure
for AI agents.
zkKYA proves who an agent is, what it can do, and how it has behaved. It gives AI systems verifiable trust without exposing the human principal behind them.
Who this is for
5
Credential types
APC, ACC, APF, AIC, ARC
First
Agent credential framework
Purpose-built for AI in Web3
Full
Principal accountability
Every agent linked to a human
Pre-tx
Intent verification
Before execution, not after
From registration to reputation — a complete
agent trust lifecycle.
Principal Registers Agent
A human principal registers their AI agent via APC, establishing a verifiable link between agent identity and the responsible human.
Capability & Payment Authorization Issued
The agent receives a Capability Credential (ACC) verifying its certified capabilities, and a Payment Facilitation Credential (APF) authorizing TEE-bound payments — both checked before the agent can operate.
Intent Declared Before Action
Before high-value actions, the agent declares intent via AIC. Principals and protocols can review before execution.
Reputation Builds Over Time
Every successful action builds the agent's Reputation Credential (ARC). Protocols and counterparties can grant progressively greater access to agents with a verified track record.
How the AI economy
uses zkMe
Select a use case to explore the problem, solution, and outcome.
Protocols cannot verify if a transaction is from a human or an AI agent, or whether the agent is authorized and accountable.
zkKYA credentials establish agent identity, capability authorization, and principal accountability before any agent transaction is accepted.
Protocols can accept agent transactions with accountable, on-chain scope enforcement.
No standard way for agents to verify each other's identity, capabilities, or trustworthiness before delegating high-value tasks.
ACC and ARC let agents verify each other's certification and reputation on-chain before collaborating.
Multi-agent systems gain verifiable trust and safer delegation.
AI trading agents must prove authorization, comply with AML, and generate audit trails — without exposing the trading strategy.
zkKYA proves the agent is authorized to trade, while KYT monitors patterns for AML compliance and supports an auditable trail.
Trading agents can operate in regulated markets while keeping strategy confidential.
Agents can execute irreversible on-chain actions without any mechanism for the principal to review intent before execution.
AIC requires intent to be declared and signed before high-value actions.
Principals keep oversight of high-risk actions with a verifiable audit trail.
The credentials that power this solution
Each credential plays a specific role. Composable — use one or combine several.
Verifies AI agent identity, authorization scope, and principal accountability. The foundational credential for the autonomous agent economy in Web3.
Real-time AML risk scoring for agent-initiated transactions. Detects anomalous patterns and ensures compliance without exposing trading strategies.
