Custom API

Creating a custom model is a process that can vary depending on the complexity of your data and the results you want. To guarantee you a solution adapted to your needs, we organize workshops to understand your data and the desired output results. This allows us to define a training schedule for your AI model based on your specifications. In general, training a custom model takes between one week and one month.

Our optical character recognition (OCR) system for documents can be trained with examples from over 50 countries using the Latin alphabet, ensuring the ability to extract data from your documents no matter where they were created. Our document recognition API is based on our computer vision technology that does not rely on text to extract data, but only on the image.

The processing time is approximately 3 seconds per document page. We continue to improve this processing time for documents and our aim is to be under 1 second. For large volumes of documents, you can use asynchronous mode for increased availability.

Dataleon offers technical support for its users. We have a dedicated team of machine learning engineers based in Paris, available from 9:00 a.m. to 6:00 p.m. (Central European Time). However, you can send us a request for assistance outside of these hours and one of our engineers will get back to you as soon as possible. The most effective way to contact our support team is by sending us an email at support@dataleon.ai.

Yes, your documents are deleted. When processing your documents or images, we create a temporary server. After processing, the document and container are removed. Your data is no longer accessible, not even for Dataleon.