Coinbase developers have implemented an innovative payment protocol x402 that allows autonomous artificial intelligence agents to make cryptocurrency payments on the Ethereum network without human involvement. The solution is based on a combination of the old web standard HTTP 402 and the proposal for Ethereum improvement 3009 (EIP-3009), which opens up new possibilities for automating financial transactions.
This is reported by Finway
Innovations for Autonomous AI Agents
Coinbase programmers Kevin Leffew and Lincoln Moore believe that autonomous AI agents will significantly expand the applications of Ethereum technologies. With x402, API services can function as cryptocurrency wallets accessible directly by artificial intelligence. This eliminates the need for key management and manual transaction confirmations, creating a foundation for fully automated economic activities.
According to the developers, such a system significantly enhances the efficiency of AI solutions and can find applications in various fields: from autonomous taxis to AI models that generate content on demand, and applications that autonomously pay for data storage.
Advantages of x402 and the Role of Ethereum
Leffew and Moore drew an analogy between the actions of AI agents and the operation of vending machines: upon receiving an HTTP 402 request, the program signs the transaction and immediately pays for the service. In the developers’ opinion, Ethereum is the best fit for such a model due to atomic payments, flexible programmable policies, and the ability to compose wallets. This significantly simplifies billing and reduces the risks of payment disputes compared to traditional systems.
“The magic of x402 lies not only in the fact that agents can pay, but also in their ability to autonomously combine different services into a single economic cycle,” emphasized Leffew and Moore.
The developers noted that testing of AI agents using HTTP 402 is already underway. As examples, they mentioned the Hyperbolic Labs project, where this technology is integrated into a large language model, and Prodia Labs, which uses it to create images and videos based on artificial intelligence.