The agentic internet is already here.

AI agents are browsing, transacting, voting, and earning on behalf of real people, and the volume of autonomous agents operating onchain is compounding fast. As that economy scales, a foundational question gets harder to ignore: when an agent shows up to transact, borrow, vote, or claim a reward, how does anyone know what's actually behind it?

ERC-8004 is Ethereum's answer to how agents trust each other. Self Protocol is the answer for everything that comes after.

What Is ERC-8004?

Proposed in August 2025 by contributors from MetaMask, the Ethereum Foundation, Google, and Coinbase, ERC-8004: Trustless Agents is an emerging Ethereum standard designed to give autonomous AI agents a verifiable identity and a portable trust layer onchain.

Before ERC-8004, two AI agents meeting for the first time had no standardized way to evaluate whether the other was reliable, capable, or safe to work with. The standard addresses this through three lightweight onchain registries. The Identity Registry gives every agent a censorship-resistant, ERC-721-based identifier that travels with it across the ecosystem. The Reputation Registry provides a standardized interface for posting and receiving feedback over time, letting agents build track records. The Validation Registry handles cryptographic verification through pluggable hooks that support ZK proofs, Trusted Execution Environments, staking, and other mechanisms for validating an agent's outputs.

What makes ERC-8004 powerful is that it's intentionally lean. Application-specific logic stays offchain, while the registries handle the primitives that matter most: who an agent is, whether it's trustworthy, and whether its work can be verified.

The Gap ERC-8004 Doesn't Address

ERC-8004 is a meaningful step for autonomous agent infrastructure, but the standard is focused on how agents trust each other. The deeper question, the one that matters enormously for protocols handling real value, is who the human behind the agent actually is, and whether that relationship can be verified at all.

When an AI agent borrows funds, claims a yield boost, purchases access to a service, or participates in governance, the protocol on the other side has a real problem. Is there a single, real human operator behind this agent, or is the activity being driven by a bot farming rewards across hundreds of wallets? Is the human principal compliant with the relevant regulations? Has an actual verified person authorized this agent, or is it just automation all the way down?

ERC-8004 gives agents an identity. Self Protocol gives those agents a verified human behind them.

Self Protocol: The Human Layer for the Agentic Web

Self is a privacy-first, ZK-powered identity protocol that lets real people prove who they are without ever exposing sensitive personal data. Through zero-knowledge proofs and biometric passport verification, users can selectively disclose their humanity, nationality, age, and other real-world attributes onchain while keeping the underlying data entirely private.

With over 11 million users and support for 129 countries, Self already powers human verification at internet scale, and as the agentic web grows more complex, Self becomes the connective tissue between autonomous agents and the verified humans authorizing them.

Where Self Fits in the ERC-8004 Stack

The overlap between Self Protocol and ERC-8004 runs deeper than it might first appear. At the Identity layer, when an agent registers onchain through ERC-8004, Self can anchor that registration to a verified human principal, giving the agent a provable, privacy-preserving link to a real person rather than an anonymous wallet. At the Validation layer, Self's ZK proofs map naturally to ERC-8004's pluggable verification hooks. Protocols that need to confirm an agent's human operator is OFAC-compliant, outside a sanctioned jurisdiction, or above a required age can call on Self's existing infrastructure to do exactly that. And at the application layer, any protocol building on ERC-8004 can use Self to gate access to verified humans, or to unlock tiered permissions based on real-world credentials like nationality, residency, or institutional membership, without collecting or storing personal data.

Already in Production Across the Agentic Stack

None of this is speculative. Self is already functioning as the human verification layer across some of the most important platforms in Web3.

Google Cloud integrated Self Protocol into its Web3 Testnet Faucets, using Self's proof-of-humanity ZK proofs to ensure test tokens reach real developers rather than bots or autonomous agents. It's one of the clearest live illustrations of the problem ERC-8004 surfaces: when resources flow onchain, distinguishing humans from agents is no longer optional.

Aave integrated Self's ZK proof-of-humanity to offer verified humans boosted yield on USDT and WETH, creating a direct financial incentive for human verification in a DeFi environment increasingly populated by autonomous actors. 

Building in the Agentic Ecosystem

The ERC-8004 era will generate an enormous volume of agent-to-agent activity, and protocols that haven't thought carefully about human verification will face Sybil attacks, bot-driven reward farming, and growing compliance exposure. Self provides the foundation that makes agentic products actually safe to deploy at scale, and it's available to integrate today.

For teams building agentic protocols or applications, Self's open-source SDK makes it straightforward to add ZK proof-of-humanity to your stack. You can verify that users are real people, check nationality and age against real-world credentials, and build Sybil-resistant reward flows without touching or storing personal data.

Start Building with Self: 

• Get In Touch with the Self core team

• Learn more about Self for AI Agents

• Browse agentic projects building with Self

• Start Exploring Integrations in the Self Playground