Ethereum Foundation Unveils zkAPI for Private AI Payments
The Ethereum Foundation and the Open Anonymity Project have launched zkAPI on Ethereum mainnet, a system for making private AI payments. This innovation addresses a pressing issue in the AI industry: the exposure of sensitive user data through account-based payment systems. AI APIs currently run through accounts tied to a card, allowing providers to link user prompts to a single profile, revealing sensitive information like health questions and financial worries.
zkAPI works by creating a vault contract on Ethereum, which is funded once with USDC. From then on, the user's device generates zero-knowledge proofs to unlock short-lived API keys, without revealing the source of the payment. This setup ensures that the payment server never sees the user's prompts, and the AI provider never sees the user's identity.
The zkAPI system is designed to be plug-and-play, allowing existing apps to switch over by pointing to the local client. There is also an in-browser option, OA Chat, and the user's balance remains withdrawable from the vault even if zkAPI's servers go dark. The system's architecture can be extended beyond chatbots to meter RPC queries, image generation, VPN bandwidth, and agent-to-agent payments with no accounts involved.
However, it's essential to note that zkAPI only hides the payment link, and providers can still see the user's IP and typing history. For stronger privacy, the team suggests routing through Tor. The launch of zkAPI is seen as a concrete step towards making AI better for users today, with the potential to become standard plumbing if AI agents transact at scale.