👁 Major updates on OCRs!

Hi everyone. In the previous months we teased you about new OCRs that are coming. Here are some fresh updates.

  • We just released in beta our own machine-learning OCR. From our extensive tests, on screens it’s much better than any other engine: Google, MS, Abbyy, etc.
    OCRs are a crucial part of ComputerVision, we’ve seen the majority of CV issues were OCR related, and this engine will finally allow you to automate without worrying about missing texts, unstable/missing anchors, and most text issues. So feel free to get it from the Beta feed in Studio and give it a try. And don’t forget to post here any feedback, it’ll help us make a better product for you.
  • We also recently released a stable version of Omnipage OCR (available in the default Studio feed). It’s a much better alternative than the default Microsoft one, so we strongly suggest you use that in production until we release a stable version of our own OCR.
  • These two join the already existing beta Abbyy Screen OCR.

All these are and will be free of charge, and the UiPath Screen OCR will be the new default OCR engine in ComputerVision once it’s released to stable.

Enjoy and tell us what you think!


@Cosin , when you say, “machine-learning OCR”, which activity are you referring to? “Machine Learning Extractor”? or other activity?

1 Like

Hi @Cosin ,

What exactly you mean by machine-learning OCR and also I cannot see Omnipage OCR in the below version,

It is [quote=“Luis_Martirena, post:3, topic:156240”]
@Cosin , when you say, “machine-learning OCR”, which activity are you referring to?

It is a new OCR activity called UiPath Screen OCR that is available in the Package Manager on the official feed, just make sure you check the Pre-release option.

1 Like

For the machine-learning OCR see the post above. For the Omnipage OCR, you need to install it in the Package Manager. Just click its icon in the main toolbar and search for the Omnipage activities package

1 Like

Got it! @Cosin . Will use it .