[Proposal] Verifiable Receipts — Proof of Exchange for A2A Economy

[junbeomlee]

Context

The x402 ecosystem is growing fast — what began as a simple payment specification is quickly becoming the foundation of the agent-to-agent (A2A) economy, enabling AI agents, APIs, and humans to transact autonomously.

As usage scales, one key limitation is emerging: the blockchain only records that money moved, but not why it moved or what happened in return.
In other words, we currently have proof of payment, but not proof of exchange — an incomplete form of trust for autonomous commerce.

Problem

In A2A interactions:

• Payments are verifiable on-chain, but the request context and result remain off-chain.
• Agents (and humans) cannot easily prove that a specific service, data, or output was actually delivered in exchange for a payment.
• This gap limits auditability, accountability, and automated dispute resolution.

Proposal: Verifiable Receipts

Introduce a verifiable receipt layer that cryptographically links the full exchange flow:
request ↔ payment ↔ result.

Each receipt could include:

{
  "version": "x402-1",
  "request_hash": "0x...",
  "payment_tx": "0x...",
  "output_hash": "0x...",
  "payer": "0x...",
  "payee": "0x...",
  "timestamp": "...",
  "signature": "..."
}

• Request hash: canonicalized body of the original request.
• Payment tx: on-chain or off-chain payment reference.
• Output hash: hash of the delivered response or file.
• Signature: from the payee (service provider) to guarantee authenticity.

[ricp]

Interesting! So the receiver of the service/data would sign the receipt when it’s delivered, but how should delayed, partial, or physical fulfillment be handled?

Physical goods introduce a time gap between payment and delivery confirmation, which can be modeled using receipt states (pending, delivered, confirmed, disputed, ...).