TheBlock
The Ethereum Foundation launched zkAPI on Thursday, a zero-knowledge system live on Ethereum mainnet that lets users pay for AI model calls and other metered APIs without revealing their identity. Built with the Open Anonymity Project, zkAPI has users deposit ETH or USDC into a vault contract; their device then generates a ZK proof that the spend is funded without revealing which note or account it draws from, and a nullifier catches double-spend attempts.
Why it matters
The design is co-authored by Ethereum co-founder Vitalik Buterin and Davide Crapis, who heads the Foundation's dAI team formed in September 2025. That team previously shipped ERC-8004, a standard for AI agent identity, in January. Together, the two releases frame a roadmap in which Ethereum functions as more than a rail for AI agents to pay other services through. The structural pitch is Ethereum as the settlement and coordination layer where AI agents transact on-chain without exposing who is behind them. Stanford PhD candidate Ken Liu of the Open Anonymity Project called zkAPI "a generalization of OA unlinkable inference that also abstracts payments away," adding that "more infrastructure should exist where privacy and sovereignty is foundational and people are in control."
Market impact
The system accepts ETH and USDC for vault deposits, routing real transaction flow through Ethereum mainnet each time a user tops up. zkAPI works with existing OpenAI and Ollama clients by exposing standard APIs on localhost, so adoption hinges on developer integration rather than building a parallel tool stack. The protocol remains experimental per the project's GitHub, and zkAPI does not provide network anonymity; users seeking stronger privacy still route through Tor, and prompt content itself can re-link sessions through personal details, writing style, or conversation history. As the first live payment rail Buterin's own work has produced for the AI-on-crypto niche, the read is bullish for the structural thesis of Ethereum as the settlement layer for AI commerce rather than short-term ETH flow.
Frequently asked questions
-
What is zkAPI and how does it work?
zkAPI is a system launched on Ethereum mainnet by the Ethereum Foundation that lets users pay for AI model calls without revealing their identity. Users deposit ETH or USDC into a vault contract, and their device generates a zero-knowledge proof showing the spend is funded without exposing which note it draws from.
-
Who co-authored the zkAPI design?
The design is co-authored by Ethereum co-founder Vitalik Buterin and Davide Crapis, who heads the Foundation's dAI team. It builds on a Feb. 11 Ethereum Research forum post titled ZK API Usage Credits.
-
What AI tools does zkAPI work with?
zkAPI exposes standard OpenAI and Ollama APIs on the user's own machine at localhost, so existing chat clients, editors, and agents can plug in directly. Use cases listed in the blog include AI chat, image and video generation, blockchain RPC queries, VPN bandwidth, and machine-to-machine payments between agents.
-
What are zkAPI's privacy limitations?
zkAPI does not provide network anonymity; the gateway can correlate requests from a stable IP address, so users seeking stronger privacy need to route through Tor. Prompt content itself can also re-link sessions through personal details, writing style, or conversation history.
-
How does zkAPI fit Ethereum's AI strategy?
zkAPI is the second major release from the Foundation's dAI team, which was formed in September 2025 and shipped the ERC-8004 standard for AI agent identity in January. Buterin's co-authorship signals the EF's near-term commitment to positioning Ethereum as the privacy and settlement layer for AI commerce.
Crypto News
WatcherGuru
CoinTelegraph
Crypto Rank News
CryptoSlate