DEVELOPER REFERENCE · V1

API errors and retry handling

Use the HTTP status and error.code together. Separate request failures from asynchronous processing failures.

Read the error object

{
  "success": false,
  "error": {
    "code": "MISSING_FILE",
    "message": "No document supplied",
    "status": 400
  }
}

Choose the next action

Status / codeAction
400 · INVALID_DOCUMENT_TYPE / MISSING_FILECorrect the document type or supply the file/URL.
401 · UNAUTHORIZEDCheck the Bearer token.
403 · FORBIDDENVerify the API key has access to the requested operation.
404 · JOB_NOT_FOUNDCheck the jobId and account context.
413 · FILE_TOO_LARGEReduce the upload size to the documented limit.
429 · RATE_LIMITEDBack off and respect Retry-After when returned.
500 · INTERNAL_ERRORRetry bounded reads with backoff; escalate repeated failures.

Do not duplicate an uncertain upload

If an upload times out after the server may have accepted it, reconcile the job before submitting again. Do not assume an idempotency header is supported unless it is in your agreed API contract. Use bounded exponential backoff with jitter for polling, and surface a timeout to the caller rather than polling forever.

Aligned with Number7 AI’s published API documentation, reviewed 8 September 2026. These examples have not been executed against your account.

Request the diagnostic