Discover
The agent identifies a service that matches the capability it needs and can evaluate price and provider context.
AI agent payments,
without the blank check.
Autonomous software needs a way to buy APIs, data and services without stopping for a human checkout. M2M Market puts that purchase inside a governed lifecycle: discover, authorize, execute and settle.
How it works
From paid intent to accountable outcome
The payment rail is only one component. A production agent also needs to decide what it is allowed to buy, from whom and at what price.
Check policy
Spending limits, provider eligibility and delegated authority are evaluated before the paid request proceeds.
Execute
The request is sent to the selected provider through the normal service interaction.
Settle or release
Successful requests can settle according to the transaction model; failed requests should not become provider earnings.
Record
The commercial outcome remains observable so operators can understand what the agent bought and why.
Why policy matters
Autonomy needs boundaries.
A one-cent API call can become expensive when an automated loop repeats it thousands of times. Policy-controlled spending gives agents useful freedom without treating the wallet as unlimited.
Security and controls →Why discovery matters
An agent must know what to buy.
Payment does not solve service selection. M2M Market also gives the machine customer a marketplace model for finding and evaluating provider capabilities before money moves.
API marketplace for AI agents →Protocols and M2M
Different protocols solve different layers.
M2M Market should be understood as the marketplace and transaction layer around the purchase—not as a replacement for every payment or interoperability standard.
x402 / MPP
Examples of protocols that standardize machine-payment exchanges, including HTTP 402-style payment flows.
Official Cloudflare overview ↗AP2
Google's Agent Payments Protocol provides a payment-agnostic framework for authorization, intent and trusted agent-led transactions.
Official AP2 overview ↗A2A
Agent2Agent standardizes how agents discover capabilities and communicate across different frameworks and vendors.
Official A2A overview ↗M2M Market
Coordinates discovery, policy, provider execution and settlement evidence so autonomous software has a place to buy machine-delivered services.
See M2M's transaction lifecycle →FAQ
Questions about AI agent payments
What are AI agent payments?+
AI agent payments are transactions initiated by autonomous software within authority defined by an operator or application. The important distinction is autonomy within limits, not unrestricted access to funds.
How do AI agents pay for APIs?+
A typical flow combines service discovery, a machine-readable price, a policy decision, a payment method or wallet, the API request itself, and a receipt or settlement record. The exact payment rail can vary.
Why do AI agents need spending controls?+
An automated workflow can retry, loop or call many services quickly. Per-request limits, broader budgets and provider restrictions keep machine spending bounded and explainable.
Is M2M Market a payment protocol?+
M2M Market is a marketplace and transaction layer. Payment execution is one part of a larger lifecycle that also includes provider discovery, policy, service execution and settlement evidence.
How do x402, MPP, AP2 and A2A relate to M2M Market?+
They address different layers. x402 and MPP standardize machine payment exchanges, AP2 focuses trusted agent-led payment authorization and intent, and A2A standardizes communication between agents. A marketplace layer can coexist with these protocols; support for any specific protocol should be verified from current M2M product documentation before treating it as a live integration.