Document Understanding - Machine Learning Extractor Dataflow Qestion

Hi All,

I’m new to DU. I want to clarify how machine learning extractor works. Currently I’m using the pre-trained invoice model of Uipath. My question is how does the extractor works, is the document being forwarded to the endpoint and the extraction happen there or the endpoint is being pulled and the extraction happen on the local studio?


The document is sent to an endpoint for the community edition and the cloud edition for extraction. However if you are using the on-prem orchestrator and Ai Frabric everything is hosted inside your network so you don’t have to worry about any data leaving or returning.

For the community and cloud edition, thus the document will be stored or it will just be extracted and will return the output?

If I have Orchestrator on-prem but havent installed AI fabric on-prem yet would that mean if I use DU with the pre-trained invoice model it would be going to the cloud or would having orchestrator on-prem be enough to keep the data from leaving and returning