Improvement

If something goes wrong while processing a document with the Klippa APIs, a meaningful message is not always returned. Instead of "something went wrong", provide the actual or a more meaningful error message.

Currently it's required to send two separate images or a PDF with two pages. It would be more convenient to also support cases where the front and back side of an ID/driver's license is on the same page.

Currently, the bank statement model on the DH Platform doesn't include a 'currency' field on header level. Add this field to the output.

Improve reading table structures on invoices or similar documents and correctly extracting and mapping their contents. Currently, when the table structure is complex, doesn't have clear headers and/or contains a lot of text, the extraction quality...

When line tems span multiple lines, the second part of the description often isn't picked up completely. Support the extraction of the additional information in the second/third line.

When listing all documents using the e-invoicing API endpoint, I always get all the documents which I received. To identify which documents are new, I would need to keep a database and check which are new documents. Instead, I would like to do a L...

I would like to receive both JSON and XML in our e-invoicing service, when receiving a document. This would be required to both show the XML to our users and use the JSON for proper data mapping in our software.

If we invite users for an organization or project after the kick-off, but they never sign up, we have no way to track whether the invite was indeed sent out or forgotten.

It would be useful if you could see invoices sent over the PEPPOL network via the DocHorizon access point, a simple UI showing the status of these invoices so you have insight in invoices Rejected because of a problem with the format because the r...

In the financial model, a hash is created for each document based on merchant name, document date and total amount. Extend the duplicate detection with a file-based hash, meaning even the slightest change in the document results in a different has...