UiPath Document OCR inside Digitize Document asks for "Image" input instead of working automatically

Hi everyone,

I’m building an Invoice/Receipt automation using REFramework + Dispatcher/Performer + Document Understanding.

My setup

  • UiPath Studio (Windows project)
  • UiPath.PDF.Activities 4.4.0-preview
  • UiPath.IntelligentOCR.Activities 7.4.0-preview

The Dispatcher adds the PDF file path to an Orchestrator Queue successfully. The Performer retrieves the QueueItem correctly.

I’m reading the file path using:


pdfPath = in_TransactionItem.SpecificContent("FilePath").ToString

Then I added a Digitize Document activity.

  • Document Path = pdfPath
  • Output:
    • documentText
    • documentObjectModel

Inside the OCR Engine section, I added UiPath Document OCR.

Problem

Instead of working automatically, UiPath Document OCR shows a mandatory Image input.

I expected Digitize Document to pass the document internally to the OCR engine, but it is asking me to manually provide an Image.

I also receive validation errors because no image is supplied.

Screenshot

(Attach the screenshot showing the Image field.)

Questions

  1. Is this expected behavior with UiPath.PDF.Activities 4.4.0-preview and IntelligentOCR 7.4.0-preview?
  2. Is this a bug in the preview packages?
  3. Should I downgrade to stable versions (PDF Activities 4.3.0 and IntelligentOCR 7.3.0)?
  4. Is there another OCR engine/activity I should be using instead of UiPath Document OCR?

Any guidance would be appreciated. Thanks!

Hi @Abhinay_Reddy1

In the UiPath Document OCR activity you have an option Use a File Resource click on that, you will be able to give the file path and use the below syntax.

LocalResource.FromPath(pdfPath)

Regards
PS Parvathy

Hi Parvathy,

Thank you for your response.

I tried using LocalResource.FromPath(pdfPath) in the Image file property, but I’m still unable to get it working.

My setup is:

  • UiPath Studio Windows project
  • UiPath.PDF.Activities 4.4.0-preview
  • UiPath.IntelligentOCR.Activities 7.4.0-preview
  • REFramework Performer
  • pdfPath contains the full PDF file path retrieved from the QueueItem (e.g. C:\...\receipt.pdf)

Could you please clarify:

  1. Should the UiPath Document OCR activity receive the PDF itself, or does it require an image resource?
  2. Is LocalResource.FromPath(pdfPath) the correct expression for a PDF file, or should I convert the PDF pages to images first?
  3. Are these package versions (PDF 4.4.0-preview and IntelligentOCR 7.4.0-preview) supported together?

Thank you!

Hi @Abhinay_Reddy1

Not needed. You can use that expression over there. It will work.

Regards
PS Parvathy

Hi @Abhinay_Reddy1

to answer your questions directly, when UiPath Document OCR is nested inside Digitize Document as the OCR engine, it shouldnt ask for a manual Image input at all, Digitize Document handles passing the document internally through its own Document Path property. The fact that Image is showing up as a mandatory field is usually a sign the activity isnt binding correctly to its parent scope, wich points more towards a package issue than something you’re doing wrong

given you are on PDF Activities 4.4.0-preview and IntelligentOCR 7.4.0-preview together, id treat that as the first suspect. Preview packages arent always tested against each other’s exact version, and a mismatch between them is a common cause of an activity’s internal scope resolution breaking like this, wich would explain the “asking for image manually” symptom exactly

to isolate it, id do what you already asked in your third question, downgrade both to the latest stable pair (PDF 4.3.0 and IntelligentOCR 7.3.0) and test the same workflow. If Image stops showing up as required there, that confirms its a preview package compatibility issue and not your workflow logic, and you can report it as a bug against the preview builds specifically

and no, you shouldnt need to convert pdf pages to images first, Digitize Document with an OCR engine attached is built to handle that conversion internally, so if it keeps asking for Image even on stable versions, that would be the real signal something’s off in how the activity was added to the sequence, might be worth deleting and re adding UiPath Document OCR as a fresh child activity instead of reusing the one thats acting up.

hope this helps narrow it down :slight_smile: