• English
• Dutch

Go to website

• How does the OCR process multi-page PDF’s?

  By default we look at the first and last page. There is a template that processes multi-page documents, activation upon use case.Few readers

• How does the Klippa OCR-API work with our system?

  The Klippa OCR and your software/system communicate with each other per means of an API. The API documentation provides the technical instructions on how the Klippa OCR connects to your software and which data points can be extracted. The API documentation can be found under: https://custom-ocr.klippa.com/docsFew readers

• What are the languages supported for the SDK?

  All the platforms that supports native elements.Few readers

• How long does it take to get the output from the API back?

  In general the time between upload and JSON response varies between 1-15 seconds. Factors that influence the output time are document size, amounts of pages, and the amount of fields that are extracted.Few readers

• How does Klippa train the OCR parsing-engine?

  We partially train the parser based on rules. That entails finding patterns by the eye and making sure that the system also recognizes the pattern. Partially it is based on Machine Learning: we tell the machine what value is correct, feeds it a big set of data and the ML-model will find the patterns among the data. We operate continuous improvement every two weeks.Few readers

• What information, if any, is stored in logs/audit trails and do the API calls prevent caching of data?

  The logs show timestamps of each API request we see, including a processing time and an API address. The API actively makes use of data caching to optimize the processing speed.Few readers

• Where are the Klippa servers located?

  By default, Klippa’s servers are hosted by Microsoft Azure and they are located in Amsterdam, Netherlands. Other hosting options are possible, namely: deployment in region of choice, dedicated hosting, private/on-premise deployment. Check https://www.klippa.com/docs/additionals for more information.Few readers

• How does the OCR connect to our system?

  Via the API. It is well documented in the following doccument: https://custom-ocr.klippa.com/docsFew readers