Skip to main content

Chargeback history

Chargeback methods expose read-only dispute history and safe evidence for your payments. Creation, evidence upload, status changes, and operational notes remain operator-managed.
Prerequisites: use an API credential with chargebacks.read. New chargeback webhook subscriptions are opt-in and are not added to existing subscriptions automatically.

Methods

All methods require chargebacks.read and are sent to POST /public/api/multihub/v1 without service_id.

Get a chargeback

Chargebacks intentionally have no merchant c_id; use the server-generated UUID directly.
Example projection:
amount is a digit-string in minor units. Reasons can be fraud, duplicate, product_not_received, customer_dispute, bank_reversal, or other.

List and summarize

chargeback.list accepts an optional payment reference, ISO 8601 created_from/created_to, limit, and cursor:
Results contain items and next_cursor; total is omitted. chargeback.summary returns:
Currency totals remain digit-strings in minor units.

Evidence attachments

List evidence metadata using the parent chargeback UUID:
Each item contains only identifiers.h_id, name, mime_type, size, and created_at. Pass its h_id to attachment.get to receive a five-minute signed download_url. Object keys, staff identities, raw metadata, internal notes, and payout details are never returned.

Error handling

Best practices

  • Store chargeback h_id and payment h_id together for reconciliation.
  • Treat webhook events as change notifications and query the current projection for details.
  • Download evidence only when needed; signed URLs expire after five minutes.
  • Never infer that 404 proves an identifier is unused globally; not-found behavior is tenant-safe.