Ethereum Foundation Launches zkAPI for Private AI Payments

on Mainnet

Ethereum | AI | Privacy | Zero-Knowledge | October 2, 2026

The Ethereum Foundation has launched zkAPI on Ethereum mainnet, a new system designed to let users pay for AI models and other metered APIs without directly linking their identity to their payments.

Built with the Open Anonymity Project, zkAPI uses zero-knowledge proofs to separate the payment trail from the user's identity. The system is currently experimental.

What Is zkAPI?

zkAPI allows users to deposit ETH, USDC or other supported assets into an Ethereum vault and receive private usage credits.

When a user wants to access an AI service, their device generates a zero-knowledge proof showing that they have sufficient funds without revealing which deposit belongs to them.

The system then provides a short-lived API key with a spending limit. AI requests can go directly to the provider, while payment settlement happens separately.

How zkAPI Works

  1. Deposit: Users fund a private balance in an Ethereum vault.
  2. Proof: The user's device generates a zero-knowledge payment proof.
  3. Authorization: zkAPI verifies the proof and issues a temporary API key.
  4. AI request: Prompts are sent to the AI provider.
  5. Settlement: Actual usage is deducted from the private balance when the session ends.

This structure is designed to prevent the payment layer from learning which user paid for a particular request.

What Does the AI Provider See?

zkAPI does not make AI usage completely anonymous.

The AI provider still receives the user's prompts and responses. Providers may also be able to observe information such as an IP address, request timing or details contained in prompts.

For example, repeatedly including personal information, project details or a distinctive writing style could potentially allow sessions to be correlated.

The Ethereum Foundation therefore describes payment privacy and network/content privacy as separate layers.

Zero-Knowledge Proofs Power the System

The protocol uses cryptographic techniques to separate payment information from user identity.

According to the Ethereum Foundation, zkAPI uses:

  • Groth16 zero-knowledge proofs
  • BN254 elliptic-curve cryptography
  • Poseidon hashing
  • A 32-level Merkle tree
  • Cryptographic nullifiers to prevent double spending

A nullifier works as a one-way identifier for a spent balance. If someone attempts to spend the same private credit twice, the duplicate nullifier can reveal the double-spending attempt without exposing the user's identity.

AI Is Only the First Use Case

Although AI payments are the headline application, zkAPI is designed for any metered API service.

Potential applications include:

Use Case Payment Unit
AI chat Model requests
AI agents Automated tasks
Blockchain RPC Queries
Image generation Individual jobs
Video generation Compute usage
VPN/bandwidth Time or data
Machine-to-machine services Per-task spending

This could be particularly relevant to autonomous AI agents that need to purchase data, computing or API access without maintaining traditional accounts.

Ethereum's Broader AI Strategy

zkAPI fits into a broader effort by the Ethereum ecosystem to develop infrastructure for AI agents.

Ethereum's dAI team has been researching agent identity, payments and coordination, including ERC-8004 and payment standards such as x402. Ethereum's own documentation describes x402 as a way for humans and AI agents to make per-use payments for web resources and APIs.

The new zkAPI system adds another layer by focusing specifically on payment privacy.

Privacy Has Limits

zkAPI does not provide complete anonymity.

The Ethereum Foundation identifies two major limitations:

Network privacy: A stable IP address or traffic pattern can potentially allow requests to be correlated.

Content privacy: Personal details, writing style or repeated conversation context can potentially identify or link sessions.

Users requiring stronger network privacy could use additional privacy tools such as Tor, while content privacy may require technologies such as local or confidential AI execution.

Why This Matters for Ethereum

The launch expands Ethereum's potential role beyond conventional crypto payments.

Instead of simply transferring tokens between wallets, Ethereum can act as a settlement layer for usage-based digital services, including AI inference and machine-to-machine transactions.

The model also addresses a growing privacy concern: traditional AI APIs often connect identity, payment information and AI usage history through a single account.

zkAPI attempts to separate those records using cryptographic proofs.

What Investors and Developers Should Watch

  • zkAPI adoption: Whether developers actually integrate the system.
  • AI agent payments: Growth of autonomous machine-to-machine commerce.
  • Stablecoin usage: Whether USDC and other digital assets become common for AI services.
  • Privacy infrastructure: Expansion of zero-knowledge applications on Ethereum.
  • x402 and ERC-8004: Development of complementary AI payment and identity standards.
  • Security testing: zkAPI remains an experimental implementation.

FAQ

What is Ethereum's zkAPI?

zkAPI is a system developed by the Ethereum Foundation and Open Anonymity Project that allows users to pay for metered APIs without directly revealing the identity behind the payment.

Can users pay AI providers with ETH?

Yes. Users can deposit assets such as ETH and USDC into the zkAPI vault and use those credits for supported API services.

Does zkAPI hide AI prompts?

No. The AI provider still receives the prompts required to provide the service. zkAPI primarily separates payment identity from API billing.

Is zkAPI completely anonymous?

No. IP addresses, traffic patterns and information contained in prompts can potentially be used to correlate activity. The project is focused on payment privacy rather than complete network anonymity.

Final Take

The Ethereum Foundation's zkAPI launch brings private, usage-based payments for AI and other APIs to Ethereum mainnet.

The technology combines Ethereum-based deposits with zero-knowledge proofs so that an API provider can receive payment without necessarily learning which deposit or identity funded it. However, zkAPI does not hide prompts, IP addresses or all other metadata.

If developers adopt the system, it could become part of a broader Ethereum-based infrastructure for AI agents, automated payments and machine-to-machine commerce. For now, the technology remains experimental and its real-world adoption will be the key test.