An error occured trying to get capabilities?

Help Document Understanding
1

Hi,

so right now I’m creating my first use case for a customer that uses AI Center and Document Understanding.

I use the model from UiPath called “Purchase Orders” and created my own version from it.

This is the current deployed version of the ML Skill (it also is public):

23.7.2.4 → so .4 is the 4th version after training.

Now in Studio I put in the asked informations like so:

image

I use the correct api key from AI Center and then this error shows as seen in the screenshot. So it doesn’t get the capabilites from it.

I already went through this thread: Machine Learning Extractor Error " An Error Occurred While Retrieving Capabilities From The Server" - News / Knowledge Base - UiPath Community Forum But wasn’t finding any issues.

We are fully in the cloud.

It also says I should look for details at the log - but which log? Where is this log?

Any ideas what could go wrong? Thx Geralt

2

Hi @ShadowDom

Are you working under enterprise network or VPN? If so, could you try in another one to check if the issue is a proxy ou firewall blocking?

3

Hi,

I indeed am working from home right now and im connected via VPN.
But then I’m connected via RDP to a VM thats inside the network, in which i develop right now.

But maybe the whole URL is blocked or something https://ai-uipath-we-b.deskover.com*
If I enter this website in the browser, I get the error that I have no permission to view this page, that should be the problem :slight_smile:

4

Hi @ShadowDom

You can’t access the URL directly from browser. It’s a normal behavior. You need call the ML Skill with right parameters and payload.

My suggestion is try to connect in other network if possible

You also can try open CMD and type the below command to check

ping ai-uipath-we-b.deskover.com and type enter

tracert ai-uipath-we-b.deskover.com and type enter

If you can’t resolve the ai-uipath-we-b.deskover.com you can check with you Cyber Team to add the ai-uipath-we-b.deskover.com in the access rules

5

Hello @ShadowDom

you can test the ML skill is accessible or not by entering the ML skill in the browser by adding a suffix “/info/model”. you will receive JSON response if the skill is accessible.

example, if ML skill URL is https://ai-uipath-we-b.deskover.com/public/mlskills/32***
then use https://ai-uipath-we-b.deskover.com/public/mlskills/32***/info/model

Cheers