Pay-per-request is a pricing and access-control model in which each individual API call or resource request is metered and paid for independently, rather than through a subscription or bulk allocation. It is enabled by protocols such as L402, which gate HTTP responses behind a Lightning Network micropayment, making per-query settlement economically viable for machine-to-machine transactions. It is particularly relevant for AI services that need frictionless, automatable payment for compute-intensive queries.