Document errors

There are different types of errors, which will be described here.

API Errors

These errors apply to cases like wrong permissions, not delivering required data etc… As a reference, you can check out the http error codes

Document errors

The next class of errors is not based on api issues but rather based on the delivered documents. In these cases no extraction (classification and/or addresses) will take place. These errors will be delivered as part of the error_codes array on the document.

Error code Notice
download_failed This is generally issued, if the download of the provided files failed.
unreadable The document is unreadable. It’s blurry, to dark, washed out, anything like that.
unprocessable The document does not comply with our deliver standards.

Download errors

In case the document includes any download errors or the download did not work at all, further details can be found in the download_errors array. Within 2 hours, downloads are retried in increasing invervals, if a download is still not available, the error download_failed will be added to the error_codes array above.

Field Content
created_at The date when a download failed
http_status The http status which has been delivered by the remote server
message The message body of the returned request

Document validation errors

The last class of errors is provided along with extracted content, but is potentially not correct or complete.

Error code Notice
not_complete This can happen, if pages are missing, or if required information for a financial receipt is missing.
payment_credentials_unclear This error will be delivered, if the API decides that the document is due for payment, but did not find any reasonable payment information. It could also be the case, that it does find payment information, but they seem not to be clear or could not be validated.
invoice_inaccurate If the API can not validate the invoice, this error will be issued. Example: The total of the invoice does not match the line items.
contains_handwritten_information If we recognize any handwritten information on the document, it is not extracted. We still deliver this as an information for further handling if necessary.
document_too_long The document is too long, and line items will not be extracted
source_file_is_encrypted The document is encrypted or write protected and cannot be signed but is further processed.
signing_pdf_failed The document cannot be signed i.e. is password protected and prohibits reading of file. The document will not be processed further.
other Any other error which is not specific.

How to cope with errors?

In any of those errors we suggest to put these invoices into a review cycle to compensate for missing information.