Using Bitskout you can now extract data from different documents like invoices, receipts, screenshots, architectural drawings, etc.

The functionality is based on matching against the template therefore if you have different visual formats (e.g. different invoice format or different screens with screenshots), then you will need to create a data extraction model per each format.

Let's create a new model for an architectural drawing:

  1. First, log in to Bitskout at

  2. Navigate to the A.I. models section and press Add. The Add model screen will appear.

3. Type the model name and small description and select Data Extraction type. Then choose Extract data from Documents from the dropdown list.

4. In the section Specify objects that you are looking for drop a template file to the drag-drop area:

5. Once the file is loaded, you will see the button to start area selection:

6. Press on the button and the full screen will open with a template and region selection option:

7. Select the regions from which you want to grab data. Once you choose the region, the system will ask to name it and choose the type:

8. Select other regions and, if needed, type words to exclude into the text box. Those words will be removed from recognition. You need to specify them using comma (word1, word2, word3,...)

9. Press Save changes and close. The system will display success message:

Note: you can adjust the template later based on the performance.

10. Specify the rest of the options as you wish and save the model by pressing Apply. Now the model is created.

The next steps are:

  • Create a Custom Output to map extracted data to your PM tool fields

  • Create a workflow to use it in recipes.

Did this answer your question?