Privacy-Focused zkAPI Payment System Launches on Ethereum Mainnet
The zkAPI platform has officially deployed to Ethereum's mainnet, transforming a previously theoretical zero-knowledge proof payment concept into a functional solution for anonymous, prepaid API services.
On Thursday, the Ethereum Foundation made public that zkAPI—a platform developed in collaboration with the Open Anonymity Project—has successfully launched on Ethereum mainnet, enabling users to compensate AI providers and other API services while maintaining billing anonymity.
The system operates by having users place funds into an Ethereum-based vault, after which they leverage zero-knowledge proofs to demonstrate sufficient credit for covering API requests while keeping their deposit identities concealed, as detailed in a statement from Vittorio Rivabella, AI coordinator at the Ethereum Foundation's dAI team.
Among the primary applications of this technology is facilitating payment for AI services without creating a traceable connection between transactions and individual users. The zkAPI platform generates temporary API keys that come with predetermined spending caps. User prompts are transmitted directly to AI service providers, while payment settlement occurs independently through a separate payment infrastructure. The development team has additionally made available a local client application and software development kit, accompanied by a browser-based AI chat implementation.
This mainnet deployment builds upon a February proposal authored by Ethereum Foundation researcher Davide Crapis and Ethereum co-founder Vitalik Buterin regarding ZK-based API usage credits. The current release transforms that theoretical framework into a fully operational mainnet deployment.
The framework was introduced as a method for providing users with private, prepaid accessibility to AI and other usage-based APIs. Nevertheless, zkAPI does not obscure the actual prompt content or network metadata from service providers, which means users may still be potentially traceable across multiple sessions via IP addresses, timing patterns or data contained within their individual requests.