> ## Documentation Index
> Fetch the complete documentation index at: https://docs.zenamu.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Errors

Every error uses the same three keys:

```json theme={null}
{
  "statusCode": 400,
  "error": "Error",
  "message": "dateFrom parameter is required (format: YYYY-MM-DD)"
}
```

* `statusCode` repeats the HTTP status.
* `error` is a short label: `Internal Server Error` for `500`, `Too Many
  Requests` for `429`, and `Error` for everything else.
* `message` explains what went wrong.

There is no machine-readable error code. Branch on the HTTP status; use
`message` for logs and for showing your team what happened.

<Note>
  On `500`, `message` is a fixed generic apology. The real cause is logged on
  Zenamu's side, so quote the timestamp and the endpoint when you
  [contact support](https://zenamu.com/contact/).
</Note>

## Handling errors

Treat `400`, `401`, `403`, and `404` as client-correctable. Retrying an
unchanged request will fail the same way — fix the request, the key, or the plan
first.

Treat `429` and `500` as transient. Back off and retry, but only where repeating
the call is safe.

<Warning>
  A `500` on a credit or pass grant is **not** safely repeatable on its own. The
  grant runs in a transaction, but the response never confirmed the outcome.
  Re-read the client's balance with
  [`GET /v1/clients`](/api/v1/reference/clients/list-clients) before granting
  again, or you may double-credit them.
</Warning>

## Partial success

The bulk endpoints process each item independently, in order. One failing item
does not roll back the rest, so the batch answers `200` even when items failed.
Each `clientId` may appear only once in a batch, which keeps result-to-request
mapping unambiguous:

```json theme={null}
{
  "statusCode": 200,
  "message": "Success",
  "data": {
    "results": [
      { "clientId": 123, "success": true, "numberOfCredits": 510, "validTo": "2026-10-01" },
      { "clientId": 456, "success": false, "error": { "message": "Client not found" } }
    ],
    "summary": {
      "total": 2,
      "succeeded": 1,
      "failed": 1,
      "failedClientIds": [456]
    }
  }
}
```

Never read a bulk `200` as "everything worked". Check `summary.failed`, then map
each result back to the request at the same array index. Inspect
`results[].error.message`, correct permanent errors, and retry only selected
failed items when the cause was fixed or transient. Do not retry the successful
items.