Using stand-alone Automation Suite on-premises, how to get the OCR Endpoint?

Help Document Understanding
#1

Hi,

If I use stand-alone Automation Suite on-premises, how can I get the OCR Endpoint? Because when I using cloud endpoint with my DU license, it said the key is invalid.

I also tried using public end point from ML Skill, and still got invalid OCR endpoint.

Please help for any suggestions.

Thanks

What is the OCR Key in UiPath Data Manager
#2

Hi @riefnaibaho
Do you just need to perform OCR on scanned pages? Or you are needing advanced data extraction? (Document Understanding)

Are you using Community license?

#3

Hi @dokumentor

I need OCR to to read IDs Card, and create model from that. So, I need to label the data in Data Manager.

I’m using enterprise license.

Thanks

#4

Hi @riefnaibaho ,

Maybe you could Check the Below Docs :

You could then Navigate to the Other Sections to learn more about OCR Deployment :
image

#5

Hi @supermanPunch ,

Yes, I already installed and deployed ML package and public ML Skill to get the end point. But when I put the endpoint in Data Manager OCR setting, it said ‘OCR endpoint is not valid’. Below is the capture.

image

Do you have any sugesstions?

#6

@riefnaibaho ,

What was the name of the UiPath OCR ML Package deployed ?

#7

@supermanPunch

it was UiPath DocumentUnderstanding and OOTB IDCards ML Package

#8

@riefnaibaho ,

The Package to be deployed first should be UiPath Document OCR ML Package.

This is the OCR required to be used as an Endpoint in the Data Manager / Data Labelling.

Have you Checked the below Doc :

#9

@supermanPunch

Where is I can get the download URL for the zip package ? My UiPath sales don’t share the download URL for that.

#10

@riefnaibaho ,

The UiPath Sales Team should be able to provide you with the links for these ML Packages. Do request them and they should be able to provide you with the ML Package links.

Maybe you would have to raise a ticket with UiPath Technical Support for this :

#11

Hi @supermanPunch

Thanks for you suggestion. After I created DocumentOCR skill and use the public endpoint, it worked in DataManager OCR setting.

1 Like