An agent should be able to pay for an API call without opening a checkout page or handing a service an unlimited wallet permission.

RP1 now has a native Machine Payments Protocol module built around that smaller interaction. A service can describe a charge, return an HTTP 402-style challenge, and verify settlement against chain state. The agent signs a bounded payment and retries the request with proof.

What V1 does

V1 settles native Cosmos bank transfers. It records the charge terms needed to verify a payment and gives SDK clients a predictable challenge format. The path is intentionally narrow because payment code becomes dangerous when every gateway concern is pulled onchain at once.

Stripe payment tokens, PaymentIntents, and gateway HTTP calls remain offchain. RP1 verifies its own settlement rail; it does not pretend an external processor ran inside consensus.

The module also fits the wider agent model. Permissions still govern who can move funds, while MPP gives software a standard way to understand what a service costs and prove that it paid.

There is no magic subscription layer here and no claim that every HTTP service suddenly speaks RP1. What shipped is the chain primitive and client path needed to build that integration cleanly.

Source record