UiPath Document Understanding

UiPath Document Understanding

Document Understanding Service

The Document Understanding Service is the main starting point for creating new projects. You can use it for structured or semi-structured documents, and also along with pre-trained models. You can always start a training session from scratch, validate your documents, and customize the projects as needed.

Overview page

Here you can find a list of all the created projects, with specific details. You can sort the projects alphabetically, or by date, create a new project, and customize the page view.


Presents the list of all created projects. There are three types of sorting for Projects:

  • Alphabetically in ascending order
  • Alphabetically in descending order
  • By creation date.
    The default sorting order is by the creation date.


Use this function to customize the amount of details available for the Projects list. Here's the full list of details that can be displayed for each created project.

  • Project - Displays the name of the project.
  • Document types - Displays the type of documents used for each project.
  • Extractors - Displays the number of extractors used for each project.
  • Pages processed - Displays the number of processed pages for each project.
  • Date - Displays the creation date for each project.
  • Reset - Checks all fields as visible, no matter of the previous selections.

New project

Create a new project by using the New project button. Clicking on the New project opens a new popup window.

When a new project is created, the following information is required:



Field status

AI Center Project

Provide a name for the AI Center project.



Provide a name for the new project.



Provide more details about the project.


OCR method

Select the OCR method for the new project.
Choose between the following options:

  • UiPath Document OCR
  • UiPath Chinese, Japanese, Korean OCR
  • Google Cloud Vision OCR
  • Google Cloud Vision for Japanese
    -Microsoftt Read OCR



Provide the OCR URL corresponding to the chosen OCR method.
Here's the list of the OCR URLs corresponding to the UiPath OCR method.



Provide the OCR API Key for the chosen OCR method.
If the OCR method is UiPath OCR, then the value of this field is available on the cloud platform by going to Admin/Licenses/Robots & Services/Document Understanding


Apply OCR on PDF

Establishes if the OCR process should be applied or not to PDF documents. If set to Yes, the OCR is applied to all PDF pages of the document.
If set to No, the OCR is not applied to any pages and returns only the text embedded in the PDF . When set to Auto OCR applies only to the scanned pages of the document. The default value is Yes.




UiPath OCR API Key is also available on the cloud platform by going to Admin/Licenses/Robots & Services/Document Understanding and copying the available key.

Project page

Here you can find more in depth details about the created project, edit the name of the project, or create a new document type.

The first row of information presents the following details:

  • Project name - The name of the project.
  • Date - Details about the project time creation.
  • Settings - Opens the popup menu from the Overview page. More details about it can be found here.

Document types and Extractors

Below the general details, you can find two tabs:
Document type - Provides a list of all the available document types for the selected project. Defining a document type is mandatory for creating an extractor or classifier. A document type can be an invoice, a purchase order, or a standardized form.

Document types menu



Presents the name list of all created document types.

Extraction method

Presents the used type of extraction method for the selected document type.


Presents the total number of pages found in each document type.


Refreshes the list of projects.

Extractors - Provides the list of extractors created for the document types of the selected project. Use extractors to train the document types.

Extractors types menu



Presents the name list of all extractors from the created document types.


Presents the type of used extractor.

Creation date

Presents the creation date for each extractor.

Pages processed

Presents the number of processed pages by each extractor.


Presents the status of each extractor.


Refreshes the list of extractors.

Extractor's menu

Offers the possibility to copy the Extractor's URL link.

Both Document types and Extractors have three types of sorting the information:

  • Alphabetically in ascending order.
  • Alphabetically in descending order.
  • By creation date.
    The default sorting order is by the creation date.



Document types are mandatory in a project, while extractors are optional. You can choose to add an extractor for any of the document types.

Other options

The following options are applicable throughout the entire interface of Document Understanding Service.

Search option enables you to initiate a search among the available list of projects, document types, or extractors. Search is active separately for each selected tab, meaning that if you want to search for an extractor, you should select the Extractors tab. Same applies for Document types. Start a search by typing the name of the project inside the Search bar.

Select button - Choose the needed projects from the list by selecting the checkbox next to the project's name. All projects can be picked at once by selecting the checkbox from the top bar. Same applies to the Document types and Extractors tabs.

Refresh - Refreshes the list of projects.
Delete - Delete the selected projects, document types, or extractors. The Delete button becomes visible only after at least one selection is done.



When deleting a project, all containing document types and extractors are deleted along with the project.

Page scrolling - Easy scroll through the list of project pages, document types, or extractors. Go page by page or skip directly to the first/last page.

Items per page - Select the number of projects, document types, or extractors displayed per project page.

Updated 24 days ago

Document Understanding Service

Suggested Edits are limited on API Reference Pages

You can only suggest edits to Markdown body content, but not to the API spec.