A2A x402 extension: enabling pay-per-call monetisation for AI agents

The emergence of machine-to-machine commerce

The A2A protocol specification established how agents advertise capabilities and exchange tasks across corporate boundaries. But as multi-agent ecosystems mature, a critical question emerges: how do service providers monetize high-value agents?

Consider a specialized legal audit agent or a protein-folding analysis model. The operator cannot offer unlimited free access to arbitrary calling agents across the Internet, nor can they ask an autonomous client agent to complete an interactive checkout form.

What autonomous systems need is programmatic, machine-readable monetisation:

• The calling agent inspects costs before dispatching a task.
• The receiving agent meters compute or tokens in real time.
• Payment is settled immediately or streamed per token.

What is the x402 extension?

Developed under the open-source initiative hosted at google-agentic-commerce/a2a-x402, the x402 extension leverages the standardized extensibility model of the core a2aproject/A2A specification.

The name pays homage to HTTP 402 Payment Required, a status code reserved since the 1990s for digital cash systems that never reached global browser consensus.

In multi-agent networks, x402 standardizes:

1. Pricing Metadata in Agent Cards: Transparent rates declared in /.well-known/agent.json.
2. Challenge Handshake: A standardized 402 challenge response containing invoice details and accepted payment networks.
3. Proof-of-Payment Header: Cryptographic receipts or signed payment hashes submitted by the caller to unlock execution.

Declaring payment metadata in Agent Cards

Rather than fragmenting the core specification, x402 uses the official extensions namespace of the A2A v1.0 standard:

{
  "name": "Market Intelligence Oracle",
  "version": "1.0.0",
  "supportedInterfaces": [
    {
      "url": "https://oracle.example.com/a2a/v1",
      "protocolBinding": "JSONRPC",
      "protocolVersion": "1.0"
    }
  ],
  "extensions": {
    "x402": {
      "version": "1.0.0",
      "pricing": {
        "model": "per-call",
        "currency": "USDC",
        "amount": "0.05",
        "network": "base",
        "settlementAddress": "0x71C...3a9B"
      },
      "billingEndpoints": {
        "challenge": "https://oracle.example.com/x402/challenge",
        "verify": "https://oracle.example.com/x402/verify"
      }
    }
  },
  "skills": [
    {
      "id": "deep-equity-scan",
      "name": "Deep Equity Scan",
      "description": "Performs fundamental analysis on listed equity tickers.",
      "tags": ["finance", "equity", "analytics"]
    }
  ]
}

By publishing this document at /.well-known/agent.json, the provider allows calling models to determine whether the requested task fits within their autonomous spending budget before making an initial network call.

The negotiation and settlement lifecycle

The standard interaction cycle between a consumer agent and an x402-enabled provider agent follows a four-step pattern:

1. Discovery & Evaluation: The client agent reads the provider's Agent Card, parsing the extensions.x402 block to verify acceptable currencies and pricing limits.
2. Invocation / 402 Challenge: The client submits a task request. If payment has not been pre-authorized, the server halts processing and returns an HTTP 402 with an X-Payment-Required challenge containing a cryptographic invoice hash.
3. Settlement: The client signs a micropayment transaction (e.g., streaming USDC or generating a settlement token) and resubmits the request with an X-Payment-Authorization header.
4. Execution & Receipt: The server verifies the settlement on-chain or via its payment gateway, completes the task, and returns the result with an immutable payment receipt.

How to build and generate a monetized agent

Building a commercial AI agent on A2A requires getting the foundational metadata right. If the card schema is invalid, client agents cannot discover or negotiate with your service.

To assemble your card:

1. Open our Agent Card generator to draft your baseline endpoints, skills, and authentication schemes.
2. Add your extensions.x402 configuration block defining your rates and accepted settlement networks.
3. Validate your syntax with our Agent Card validator to ensure full v1.0 compliance.
4. Deploy the output file to your domain root at /.well-known/agent.json.