Token Billing

Token Billing is a Stripe feature that automatically measures how much an AI company's customers actually use, and turns that into an invoice — so companies can charge based on real usage instead of a flat fee. Think of it like an electricity meter: just as a meter reads how much power a household used and prints a bill for that exact amount, Token Billing counts the units an AI consumes when generating a response (tokens, the small chunks that text is broken into for processing) and converts that count directly into a charge.

Because usage varies so much from one AI user to the next, charging based on actual consumption often makes more sense than charging everyone the same flat monthly fee. The catch is that manually calculating usage and building it into invoices is tedious — Token Billing automates that process so development teams don't have to build their own billing logic from scratch.

Stripe has been building Token Billing into a connective layer between AI infrastructure and payments, having acquired usage-based billing company Metronome as well as OpenRouter, a gateway service that lets users pick from multiple AI models in one place. In fact, OpenRouter and Stripe have worked together to build integrations for this feature.

The article states: "They also launched Token Billing, which automatically meters LLM token usage and charges for it." Here, Token Billing refers not to a person manually calculating usage and creating invoices, but to a system that automatically measures usage and applies charges on its own.