The US-based payments company Stripe has introduced a payment system purpose-built for AI agents, allowing automated software to pay for digital services using USD Coin (USDC) stablecoins on the Base blockchain network. The feature, currently available in preview, was revealed on 11 February 2025 by a Stripe product manager.
The system adds support for the x402 payment protocol, an open standard that repurposes the long-dormant HTTP 402 ('Payment Required') status code. Under this protocol, when an AI agent attempts to access a paid service, it receives a structured payment request. Once the agent sends USDC on Base, access is granted automatically, without human intervention.
How the system works
From a technical standpoint, the process begins when a business creates a standard Stripe Payment Intent. Stripe then assigns a unique wallet address to each transaction. When an AI agent transfers funds to that address, the payment can be tracked in real time through the Stripe dashboard, via webhooks, or through the API. Once confirmed, the funds settle into the merchant's Stripe balance in the same way as any conventional payment. Stripe's existing infrastructure handles tax reporting, refunds, and compliance.
The choice of Base, an Ethereum-based network, is significant in this context. The network offers fast settlement times and low transaction fees, making it suitable for frequent, low-value payments such as per-request or per-minute pricing models. This contrasts with traditional subscription or prepaid credit structures, which are less flexible for automated, high-frequency transactions.
According to Stripe, current payment systems are designed primarily for human users and are poorly suited to the needs of autonomous software. AI agents require payment rails that are fast, low-cost, and available at all times without manual oversight.
Positioning within the agent economy
The launch forms part of Stripe's broader strategy around what it describes as the 'agent economy', an emerging model in which software programmes operate independently and manage their own financial transactions. In this model, agents are expected to purchase data, computing resources, and digital services without requiring human approval.
To support developer adoption, Stripe has released an open-source command-line tool called 'purl', along with sample code in Python and Node.js, designed to help developers test machine-to-machine payments.
Stripe has indicated that additional protocols, currencies, and blockchain networks will be supported in the future, though the current scope is limited to USDC on Base. The move signals a broader convergence of AI, fintech, and cryptocurrency infrastructure, with usage-based pricing models replacing fixed billing in an increasing number of machine-driven contexts.
