Pay Per Use: when AI uses your work, you should get paid

Cloudflare announced that Pay Per Use is now in beta, a system where AI companies pay publishers for specific downstream uses of their content instead of for crawling. Cloudflare first outlined the plan in July and has since worked with buyers and content owners to build it. The company also said its Monetization Gateway, which lets sellers charge agents per request using the open x402 protocol, is in beta as of today. Both products share identity, metering, pricing, and analytics infrastructure.

In Pay Per Use, an AI company sets up a program with Cloudflare by identifying its crawler, defining which use it will pay for, setting a price, and connecting a payment account. Publishers review those offers in the Cloudflare dashboard under Monetize then decide whether to accept, and can stop participating later. There is no origin change or technical integration required with each AI company. Each program's terms define what the AI company may do with the content, including any restrictions on training.

Buyers fetch the list of domains that accepted their offer, then report each use as one line of JSON containing the time, the source URL, and an event ID. Usage is self-reported, but program terms require complete reporting and Cloudflare checks each reported use against an enrolled publisher. Cloudflare aggregates the reports, charges the buyer, and pays publishers monthly through a connected payment account. Buyers need only one integration, and publishers get a single place to see offers and earnings.

One distinction Cloudflare makes is that Pay Per Crawl, launched in 2025, charges for access, while Pay Per Use pays for what happens after, such as an excerpt shown in a search answer or a review shaping a shopping recommendation. The same article can be paid for differently in different products, and publishers can accept different offers for different uses. During the beta Cloudflare is also working with AI companies to report more context about each use, such as the keywords that led to a citation or the product it powered, while stating that no personal data is exchanged.

Why it matters: Publishers get a way to accept or decline AI use of their content with tracked usage and monthly payments, while AI companies get a single API connecting them to every participating site, avoiding thousands of separate integrations.