Konfuzio SDKs#

Data extraction and classification is easy to implement and maintain with the Konfuzio SDKs.

Are you a developer?

Have a look at dev.konfuzio.com to review the technical details.

Fast OCR processing for documents#

Fast OCR in your applications is possible with the Konfuzio SDK and Trainer Package. Classify and extract receipts, invoices, contracts and passports, by either sending them to our cloud OCR API or embed the models in your software to receive structured data within a few seconds.

Data mining, extraction and classification of documents using Python or the API#

Two SDKs are available to process documents in Python. Your developers get’s you up and running within a day!

What is the difference between the free OCR SDK and the Konfuzio Trainer?#

There are two main SDKs that Konfuzio offers, one is for free and under MIT license. Our default scanning SDK can be used to scan documents and extract them via REST API, which is called Python SDK. If you are looking for a Software Development Kit that will allow you to customize the full Document AI pipeline, you want to license the Developer Package called Konfuzio Trainer.

Konfuzio SDK

Konfuzio Trainer

Document AI

via API


AI Evaluation

via API


Export Data



Export Model





developer license

Konfuzio Python SDK utilize OCR, document extraction and document classification via API#

It is possible to link the Konfuzio Python SDK to Konfuzio’s cloud based OCR API, the so-called Konfuzio Server. After scanning, a request it is then handled through the API and the JSON response is returned to your application. In some cases it is also possible to use on device OCR. This is a custom addition, so in that case a serious extra investment must be taken into account. In contrast, to on-device OCR our API based SDKs are available for free. Visit dev.konfuzio.com to find out how to install and use the free Python SDK.

Konfuzio Trainer allows users to embed document extraction and document classification in products#

The cost of our Trainer consists of an annual license fee, plus an hourly rate for any support needed in implementing it. Since costs vary per project, we are happy to discuss this with you and provide you with a quote specifically for your project. As an indication you can assume that the basic licensing costs are between €5,000 and €10,000 per year.

Further information concerning customized NER, NLP or object detection models#

Besides, the SDKs, we also offer the possibility to deliver customized NER, NLP or mask detection deep computer vision models. In the latter case you can use the model to recognize certain objects based on object detection. In addition to documents, this can be objects, for example license plates, cars, price tags, displays, product packaging or other objects. It is possible to train a custom object detection model that can be easily and dynamically accessed via the SDK. Please contact us for more information about our customized models.