Digitize Document One Or More Errors Occurred When Using UiPath DocumentOCR Deployed Skill

How to troubleshoot UiPath Document OCR Digitize Document: One or more errors occurred?

***This article's troubleshooting steps only apply if using an endpoint for an ML Skill that has been deployed with the DocumentOCR model in the customer's instance of AI Center. These steps are not applicable for troubleshooting issues when using UiPath's Public Endpoint for UiPath Document OCR.***


When attempting to Digitize a document using an endpoint for UiPath Document OCR that was deployed in the customer's instance AI Center (on-prem or cloud), the following error is seen. -

  • OCR Digitize Document: One or more errors occurred.

Troubleshooting Steps:

  • Validate that the UiPath Document OCR skill is available in AI Center.
    1. Navigate to ML Skills ->
    2. Look for the skill that has been deployed using the UiPath Document OCR model ->
    3. Validate that the skill's status is showing as "available". If the skill is not available, try resuming/redeploying the skill.

  • Validate that the UiPath Document OCR skill has been made public and the endpoint can be reached.
    1. Navigate to the ML Skill
    2. If the ML Skill has not been made public, modify the deployment to make the ML Skill public. For how to make an ML Skill public, see Make ML Skill Public/Private.
    3. After confirming the ML Skill has been made public, copy the URL for the skill.
    4. On the machine where the issue is occurring, open a browser and paste the ML Skill's public endpoint in the navigation bar.
    5. The expected message in the browser is 405: Method is not allowed. <- This is expected (If something other than this is displayed, there could be an issue with reaching the ML Skill from the machine) Capture a screenshot to share the results with UiPath Product Support later if needed. Additional troubleshooting may be required to diagnose if this is the case.

  • Validate that the correct endpoint and API key (API key not applicable for on-prem air-gapped installations) are being used in the UiPath Document OCR engine activity in Studio.

***The following step only applies for On-Prem AiCenter installations. This is not applicable for UiPath Cloud AiCenter***

  • Validate that the version of the ML Skill deployed is compatible with the On-Prem version of AI Center that is being used.
    1. Click on ML Packages
    2. Click on the Package for the DocumentOCR
    3. After the page for the DocumentOCR ML Package opens, take a screenshot of the version that is currently being used.
    4. Click Upload new version in the top right corner
    5. On this page, scroll to the bottom and look for the "Minimum AI Fabric version" listed.
    6. The version of AI Center that is currently in use is visible in the AI Center Navigation Menu.
    7. Now refer back to the screenshot that was previously taken of the ML Package version. If the version that is deployed is higher than minimum AI Fabric version that is currently deployed, a new ML Package will need to be deployed that meets the minimum requirement for the AI Fabric version deployed.

After following the Troubleshooting steps as provided above, if the issue is still occurring, please contact UiPath Support and share the following information after replicating the issue:

  1. Robot Execution Logs from a machine where the issue is occurring (The robot logs are located here: %localappdata%/UiPath/Logs)
  2. Event Viewer Logs from the same machine during the same time frame for the Robot Execution Logs. How to get Application Logs from the Event Viewer.
  3. Screenshots that were captured during Troubleshooting steps as described above.
  4. Share a screenshot of the UiPath Document OCR activity's configuration
  5. Share a screenshot of the project's dependencies, or share the project files (Xamls, project.json, etc) in a zipped folder.
  6. If AI Center is on Prem, please also share a screenshot of the version of AI Center version that is deployed and also specify if the installation is Air-Gapped.