Document Processing 20.4 Beta Returns: Robot Reading Your Handwriting

Did you think this was it? Well… Surprise!


:fireworks: :loudspeaker: :sparkler: We are happy to announce that we have released a new (yeah, even newer) beta version of the IntelligentOCR package, part of UiPath Document Understanding , that allows your robot to process Forms in a much nicer way!

The activities with shiny new features included in this beta package are the already-known Form Extractor, which gets improved Checkbox Detection, and the brand new Intelligent Form Extractor (the Form Extractor’s enhanced version), which gets you Signature Detection and Handwriting Recognition.

Form Extractor :older_adult:

You will notice just some “underground” changes in the new version of the Form Extractor.

We have improved Boolean field interpretation and added a better checkbox detection algorithm for the extractor, so you should see even better results there. Small bug fixes were also included in this release.

But the coolest part is that Form Extractor is now a daddy, as the Intelligent Form Extractor was born! :birthday:

Intelligent Form Extractor :child:

And here’s the new kid on the block, the Intelligent Form Extractor :tada: :sparkles: :crown:

You already know how to use the Form Extractor, right? Then this is going to be child’s play to set up.

Once you open the Template Manager (from the Manage Templates link in the activity), once you define a template, you will notice a couple of extra options available for you:

  • You can select, from your document type fields, which ones should be checked whether they contain a signature or not. Pretty cool, isn’t it? :ice_cube: If you define a field as Boolean in your taxonomy, and you configure it as a Signature Field, the Intelligent Form Extractor will try to determine if, in the template defined area, there is a signature or not, and report a correct Yes/No value corresponding to the found content.
  • You can select, from your document type fields, which ones you expect to have handwritten content. (Yep, this is the cherry-on-top feature right here :cherries: :cake:) The Intelligent Form Extractor knows how to read handwriting. All you need to do is tick the boxes next to the fields that are usually handwritten, and the extractor does the rest!

:exclamation: Do note that you can currently select at most 60 fields (cells included and counted separately according to template definitions) for handwriting processing on a single page.

Form Extractor and Intelligent Form Extractor - Usage Changes

You will notice that both activities now require you to enter your Document Understanding API Key from your Cloud Platform account. Both activities perform HTTP calls to their respective endpoints for authentication and processing.

:grey_exclamation: Make sure your Robots are connected to the Internet for the activities to work properly.

Community Usage Limitations

  • You can use the Form Extractor with documents of at most 2 pages, with a call limit of 50 calls/hour.
  • You can use the Intelligent Form Extractor with documents of at most 1 page, at most 5 handwritten fields, and a call limit of 50 calls/hour.

:exclamation: Exceeding the call limit will lead to your calls being blocked for an hour.

Ready to Give it a Try? :writing_hand:

Then grab :open_hands: the latest IntelligentOCR package (make sure to check the “Include Pre-release” flag in Package Manager), and start automating the processing of handwritten documents!

Make sure to send us feedback, it helps us improve the Document Understanding capabilities with every new release, and ensure that they work for you and your use cases.

Stay happy and stay safe,

The Document Understanding Team


Thank You so much. This is a great feature. Actually was waiting for that feature where handwritten documents could be extracted. My query is, it is free of cost or any charges for intelligentOCR ?


Hi @Ioana_Gligan

The Intelligent Form Extractor is pointing to “”, can we use the same key that we use for Document Understanding?



1 Like

hello @hemal,

the use of the IntelligentOCR package is free of charge. As a community user, there are some limitations that apply, just like for the Machine Learning extractor:

  • 50 documents per hour, at most 1 page and at most 5 handwritten fields for Intelligent Form Extractor,
  • 50 documents per hour, at most 2 pages for Form Extractor.

Enterprise users can use these features without any limitation to size or number of calls per hour. You will need to discuss with your Sales representative about Document Understanding bundles, to enable your enterprise to use these extractors.


Hello @AndresTarazona,

yes, it is the same key across all the document understanding functionalities, the one from your Platform account.


Thank you so much @Ioana_Gligan :slight_smile: :slight_smile: