Question 1

How do refunds work?

Accepted Answer

A refund attaches to the invoice the customer already paid and returns value toward the original payer. Because an on-chain payment is final and cannot be reversed at the protocol level, a refund is a new outbound transfer an operator approves and sends — reasoned about against the amount paid on that invoice, so it can be the full amount or part of it. Network fees and on-chain rates apply to sending it, and every refund is recorded against its source invoice.

Question 2

Can I issue a partial refund?

Accepted Answer

Yes. A refund is reasoned about relative to the amount paid on the original invoice, so you can return the full paid amount or part of it. That covers a full cancellation, a partial credit when one line item is unavailable, and returning the surplus when a customer overpaid the invoice.

Question 3

Who approves a refund?

Accepted Answer

An operator with the right permission. Refunds run through operator approval and sit behind scoped API keys and permissions, the same model that governs payouts — a read-only reporting key cannot trigger a refund. Every refund is recorded with its actor, amount, destination, and the invoice it refunds.

How do refunds work?

Keep reading across the cluster.