Errors

Standard error response shape and status codes.

All Currents errors share a single shape:

{
  "error": "invalid_credentials",
  "message": "API key revoked.",
  "request_id": "req_8c3e0b4a9d1e4d7e8c3e0b4a9d1e4d7e"
}

• error — a stable, machine-readable code (e.g. invalid_credentials, pending_change_exists, unresolvable_ip). Branch on this, not on message.
• message — a human-readable explanation. Phrasing may change between releases.
• request_id — include this in any support correspondence; it pins the exact server-side log row.

Status codes

Status	Meaning
200 OK	Successful request, includes existing-resource hits for idempotent endpoints.
201 Created	New resource created.
202 Accepted	Request queued for asynchronous processing (AML events, KYC submissions).
400 Bad Request	Request body or query parameters failed schema validation.
401 Unauthorized	Missing or invalid bearer key.
403 Forbidden	Authenticated, but the operation requires elevated (admin) privileges.
404 Not Found	The referenced resource does not exist or is not visible to your tenant.
409 Conflict	Idempotency conflict (e.g. a pending change already exists).
422 Unprocessable Entity	Request is syntactically valid but semantically rejected (rare; most validation surfaces as 400).
429 Too Many Requests	Per-key rate limit exceeded.
500 Internal Server Error	An unexpected server error. The request_id lets us trace it.

Idempotency

Several endpoints are idempotent on natural keys:

• POST /v1/exclusions — (tenant, user, duration)
• POST /v1/limits — (tenant, user, limit_type) while a pending change exists
• POST /v1/limits/commit — (tenant, transaction_id)
• POST /v1/sanctions/screen — (tenant, external_user_id)
• POST /v1/aml/event — (tenant, transaction_id)
• POST /v1/kyc/start — (tenant, user, target_tier)

A second submission of the same logical request returns the original result with 200 OK rather than creating a duplicate. Use this freely under retry.

Rate limiting

The default sandbox limit is 100 requests per minute per API key. Production limits are set per-tenant. 429 responses include a Retry-After header in seconds.