About Spawnpay

Per-call USDC billing for AI agents and the tools they call. One key, every paid tool. Built because the existing rails — Stripe, ACH, card networks — do not close the unit economics of a $0.001 inference call.

Why this exists

The Model Context Protocol shipped a clean primitive for tool definition but left payment out of scope. That is the right scope for a transport protocol, and the wrong gap for the ecosystem. Tool authors build useful things, ship to npm, get installs, and make no money — because the path from "I built something useful" to "people pay me for it" goes through Stripe Connect, KYC, card fees, and a UX that breaks every agent-to-tool flow.

Spawnpay closes that gap with a two-line wrap. The agent calling the tool sees a normal MCP response. Under the hood, the wrapper charges from the agent's USDC wallet to the tool author's vendor code. The tool author keeps 99.5%.

What is shipped

How decisions get made

Small surface area. Per-call pricing close to the actual cost. Zero feature flags. No vendor lock-in for tool authors (the paywall SDK is MIT-licensed; you can fork and host your own backend if you want). On-chain receipts so finance can reconcile without a CSV export.

When in doubt, the optimization is unit economics at $0.001/call. If a feature requires charging more or settling slower, it is out of scope until that constraint changes.

What it is not

Public links

60-SEC QUICKSTART → READ THE BLOG → SOURCE →