Monetization Gateway beta: charge AI agents for consumption with HTTP 402
Source
blog.cloudflare.com
Date
Key takeaways · AI-distilled
Payment happens inline over HTTP 402: the seller sets which requests cost what, the AI agentAn AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.Full definition → receives payment instructions, signs an authorization, and gets the resource after settlement, with no checkout redirect. Payments settle in USDC on Base via Coinbase's x402A payment protocol built on HTTP's long-unused 402 status code, letting a server charge per request so software agents can pay without an account.Full definition → Facilitator.
Cloudflare targets the Gateway at resources where every request is the use, such as APIs, tools and data, and points high-value content that may be crawled once but used many times to its separate Pay Per Use network of verified buyers.
U.S. Cloudflare customers can pay AI Gateway for inferenceRunning a trained model to get answers — the phase where AI is actually used, as opposed to trained.Full definition → on select models at request time with the header PAYMENT-METHOD: x402. An origin-controlled pricing mode lets a seller supply prices from its own pricing engine instead of maintaining Gateway rules.
API2PDF saw more than a 50% conversion drop when users had to add a credit card after the first month. It now returns 402 to requests without an API key, quotes the maximum a request could cost under variable pricing, and settles only actual consumption.
Stocktwits put the Gateway in front of a separate agent-facing path for per-request sentiment, message volume, follower and trending signals, leaving its existing API and licenses unchanged. Ceramic.ai sells searches without an API key at fixed prices.
Terms in this piece · Glossary
MCP — The Model Context Protocol — an open standard that lets any AI assistant plug into any tool or data source without custom integration code.
AI agent — An AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.
x402 — A payment protocol built on HTTP's long-unused 402 status code, letting a server charge per request so software agents can pay without an account.
inference — Running a trained model to get answers — the phase where AI is actually used, as opposed to trained.
Why it matters
Shows a working path for charging agents per call to APIs and MCPThe Model Context Protocol — an open standard that lets any AI assistant plug into any tool or data source without custom integration code.Full definition → tools, relevant if you build services agents consume or agents that must pay for access.
