Agent Principal Credential (APC)
Cryptographically binds an AI agent to its human or legal entity principal, establishing irrefutable accountability so all autonomous actions trace back to a verified real-world entity.
User Journey
Frank runs a growing startup and uses an AI agent to manage all of his company's SaaS subscriptions, from cloud hosting to design tools to analytics platforms. Each provider requires verified identity before allowing automated billing and contract changes. Frank completes a one-time zkKYC verification through zkMe and receives a private identity credential. He then uses the zkMe Vault to issue an APC to his management agent, cryptographically binding it to his verified business entity. When the agent signs up for a new service or renews an existing subscription, it presents a Zero-Knowledge Proof of its APC. The provider instantly confirms the agent is authorized by a legitimate, verified company without ever seeing Frank's personal details or corporate documents. Frank retains full control and can revoke the agent's APC or adjust its permissions at any time.
See It in Action
COMING SOON
The Accountability Gap in the Agent Economy
Traditional AI systems often operate in an accountability vacuum, creating significant risks for users, platforms, and regulators:
Unknown Owners: Agents are often deployed without a clear, verifiable link to the person or entity responsible for their actions.
Jurisdictional Arbitrage: Malicious operators can hide behind layers of anonymity and cross-border complexities to evade responsibility.
Regulatory Evasion: Without a mechanism to enforce compliance across decentralized networks, agents can be used to bypass critical legal and financial regulations.
The APC is designed to close this gap by establishing a foundational layer of trust and accountability for the entire agent ecosystem.
Why zkMe APC?
zkMe provides a privacy-preserving, technically robust, and interoperable solution for agent accountability.
Privacy & Compliance
Selective Disclosure
Principals can prove their verified status using Zero-Knowledge Proofs (ZKPs) without revealing underlying personal or corporate data, ensuring AML/KYC compliance while preserving privacy.
Technical Excellence
Battle-Tested Cryptography
Scalable & Decentralized
The system leverages production-ready implementations of BBS+ signatures for selective disclosure and Groth16 ZK-SNARKs for efficient, verifiable proofs. Built on a decentralized architecture with no single point of failure, the infrastructure is designed to handle millions of verifications with sub-second latency.
Ecosystem Ready
Seamless Integration
Interoperable Standards
The entire framework is built on W3C standards for Decentralized Identifiers (DIDs) and Verifiable Credentials (VCs), ensuring compatibility with existing identity infrastructure. Provides plug-and-play SDKs for major development frameworks and multi-chain support (including Ethereum, Polygon, and Solana) for broad adoption.
How It Works
The process is divided into a flow for principals (agent owners) and verifiers (service providers).
For Agent Principals / Owners:
Registration & Verification: The principal creates a Decentralized Identifier (DID) and undergoes a zkKYC or zkKYB verification process, receiving a verifiable credential (SBT) in their wallet upon success.
Credential Issuance: The principal requests an APC from the zkMe protocol.
Agent Binding: The issued APC is cryptographically linked to the agent's DID, creating a verifiable and tamper-proof bond between the agent and its owner.
Lifecycle Management: The principal can update or revoke the APC as ownership or permissions change over time.
For Agent Verifiers:
Request Proof: The verifier's platform (e.g., a DeFi protocol) requests proof of principal from an interacting agent.
Verify Credential: The agent provides a ZKP, which the verifier validates against on-chain state roots. This process confirms the agent has a verified principal without exposing any of the principal's sensitive data.
Risk Assessment: Based on the verified relationship, the verifier can make an informed trust decision and grant access.
Audit Trail: An immutable record of the verification is logged for regulatory and security purposes.
Credential Structure
The APC schema is designed to be comprehensive and flexible.
principal_did
The DID of the verified human or legal entity owner.
agent_did
The DID of the AI agent.
relationship_type
The nature of the relationship (e.g., owner, operator, developer).
permission_scope
An embedded object defining specific limitations (can also be a separate ASC).
validity_period
The issuance and expiration dates of the credential.
Use Cases
Financial Services
DeFi Protocols: Verify trading bot ownership before granting API access.
Lending Platforms: Assess principal credibility for agent-originated loans.
Payment Systems: Ensure AML/CFT compliance for autonomous payment agents.
Enterprise
Corporate AI Systems: Establish clear internal accountability for AI assistants.
Supply Chain: Verify authorized trading and logistics agents in a B2B network.
Data Access Control: Ensure only properly owned agents can access sensitive corporate data.
Regulatory & Consumer
Market Surveillance: Allow regulators to monitor agent ownership patterns.
Consumer Protection: Provide clear recourse paths for damages caused by misbehaving agents.
Content Attribution: Attribute AI-generated content to a responsible entity.
Technical Foundation
BBS+ Signatures: Enable selective disclosure of credential attributes, so an agent can prove ownership without revealing the owner's full identity.
Zero-Knowledge Proofs (Groth16): Allow for efficient verification of compliance claims without revealing the underlying data.
Merkle Tree State Roots: On-chain state roots provide an efficient and globally verifiable source of truth for credential status.
Revocation Registries: Ensure that credentials can be immediately invalidated if compromised or if ownership changes.
Last updated