Ethereum’s zkAPI system has gone live on the mainnet, bringing a privacy-focused way to pay for AI and other metered API services. The project allows users to fund API access in advance while limiting the amount of payment information revealed to service providers.
According to Cointelegraph, the Ethereum Foundation announced the launch on Thursday in collaboration with the Open Anonymity Project. The system is designed to separate API usage from a user’s billing identity, using zero-knowledge proofs to confirm that an account has enough credit for a request without disclosing which deposit belongs to that user.
How zkAPI works
Users first deposit funds into an Ethereum-based vault. When they make an API request, zkAPI uses a zero-knowledge proof to demonstrate that sufficient prepaid credit is available. Zero-knowledge technology allows one party to verify information without receiving the underlying details—in this case, helping confirm payment capacity without exposing the user’s specific deposit.
The system creates short-lived API keys with preset spending limits. Prompts are sent directly to the AI provider, while the related payment settlement is handled separately through zkAPI’s payment layer. This arrangement is intended to reduce the direct connection between a person using an AI service and the payment account funding that use.
The release includes a local client and a software development kit for developers. It also includes a browser-based AI chat implementation, giving users an immediately accessible example of how the payment system can be used.
Privacy has limits
zkAPI is intended to provide privacy around payment and billing information, but it does not make every aspect of an API session anonymous. Providers can still see prompt contents and network metadata. The source material says users could potentially be associated across different sessions through details such as IP addresses, timing patterns or information included in their requests.
That distinction is important: the system separates prepaid payment credentials from API access, but it does not conceal the requests themselves from the provider handling them. Its stated use case is therefore private payment for AI and other metered services, rather than complete anonymity for online activity.
From proposal to mainnet implementation
The launch follows a proposal published in February by Ethereum Foundation researcher Davide Crapis and Ethereum co-founder Vitalik Buterin. That proposal described the use of zero-knowledge technology to create private, prepaid credits for API usage.
The mainnet release turns that earlier design into an operating implementation on Ethereum. It could be used for AI services as well as other APIs that charge according to consumption, although the source material does not provide transaction volumes, user numbers or details about commercial adoption.
Ethereum market picture
Moneyiar market data at 09:18 on October 3, 2026, shows Ethereum (ETH) trading at $2,684.9728052892, down 2.44% over 24 hours and up 0.14% over seven days. Other tracked assets were mixed: Shiba Inu (SHIB) was at $0.0000056604, down 4.95% in 24 hours and 3.41% over seven days; Uniswap (UNI) stood at $9.2592243014, up 2.27% in 24 hours but down 3.75% over seven days; and Pepe (PEPE) was $0.0000042445, down 7.15% and 3.58% over those periods, respectively, according to Moneyiar data at 09:19 on October 3, 2026.
