The Ethereum Foundation and the Open Anonymity Project have launched zkAPI on Ethereum mainnet, a protocol that lets users pay for AI services without tying their identity to their requests, according to the Foundation's announcement on October 1. Users can prepay in USDC and receive capped, short-lived API keys issued through zero-knowledge proofs.

The design follows a February 2026 proposal on Ethereum Research by Davide Crapis, who leads the Foundation's dAI team, and Vitalik Buterin. The project's repository labels the protocol experimental, and its documentation identifies OpenRouter as the AI provider behind the keys.

“Prompts are personal. People ask AI models about their health, their finances, their doubts. Under the current model, using AI means handing a running transcript of your thinking to whoever holds the billing relationship,” the Ethereum Foundation argued in the announcement.

“zkAPI separates payment from identity,” the Foundation continued. “You deposit credits (ETH, USDC, etc.) into a vault contract on Ethereum, one ordinary transaction. From then on, your balance exists as a private note: digital cash that only you can spend and that nobody can trace back to the deposit.”

The system was built with the Open Anonymity Project. Alongside the launch comes OA Chat, a browser-based private chatbot that requires no installation, and the code is open on GitHub, with further details on the project's official site.

The privacy concern is not abstract. In May 2025, a federal court ordered OpenAI to preserve its output logs, including chats users had deleted, in a copyright lawsuit brought by The New York Times and other publishers.

Under the implementation, software on a user's device builds a zero-knowledge proof each time the user wants to ask an AI something—a mathematical receipt that establishes a fact without revealing the details behind it. This particular proof demonstrates that the user holds sufficient funds and has not already spent them.

The server checks the proof and hands back a temporary API key, essentially a password that lets software talk to an AI service. The key is capped at a dollar amount and lives only in the device's memory. Requests then go to the AI provider carrying that key and nothing else: no name, no payment details.

When the key expires, the provider issues a signed usage receipt, and the user is charged for what they actually used rather than the cap. The provider sees what was asked, the payment layer sees what was spent, and neither can link the two.

The project's documentation describes short-lived keys that work with OpenRouter, a service that provides access to hundreds of AI models through a single connection. Any application that speaks the OpenAI format can plug in by pointing at a local address.

The concept originated in the February proposal Crapis and Buterin posted to Ethereum Research. Crapis leads the Foundation's dAI team, which aims to make Ethereum the preferred settlement and coordination layer for AI. He had already pitched Ethereum as plumbing for software agents: in August 2025, Crapis and Marco De Rossi proposed the ERC-8004 standard, a shared rulebook that lets autonomous AI agents find each other, verify identity, and transact. Crapis predicted that most Ethereum traffic would come from machines within three to five years.\nButerin has run a parallel thread. On February 10, he laid out an Ethereum-AI roadmap that included zero-knowledge payments for private API use. In April, he demonstrated the other path: running an open-source model locally on his own Nvidia 5090 graphics card instead of feeding his life to cloud services. Local models keep data at home; zkAPI covers the case where prompts still travel to someone else's servers.

The same plumbing could cover other metered services, per the announcement: blockchain data queries, image and video generation, VPNs, and payments between machines. Providers integrating directly would have to accept a proof instead of an API key and settle signed usage receipts. So far, OpenRouter is the only provider the documentation names, so how many services adopt that direct-integration route will determine how widely the private-payment option is available beyond the launch.

The repository itself labels the protocol experimental and does not list a formal audit—a detail that carries weight for a system whose vault contract holds users' real deposits on mainnet.