Manage OCR Models

Configure Docling and Azure Doc Intelligence engines for the Doc Intelligence Node, including deployment, resource allocation, and connection management.

Configure the document-processing engines that power the Doc Intelligence Node. Supported engines—Docling and Azure Doc Intelligence—enable text extraction, layout understanding, and structured data output.

Each engine must be configured on this page before it appears as an option in the Doc Intelligence Node.

Docling

Docling provides AI-powered document extraction for text, layout, and structured data. You can create multiple deployments.

Deploy a model

1. Go to Settings → Manage Models → OCR.
2. Click Deploy a Model.
3. Enter a unique deployment name and an optional description.
4. Select a language. English (en) is selected by default. You can add one additional language to improve model performance.
5. Configure resource allocation:
   • Minimum and maximum replicas.
   • Auto-scaling threshold.
   • Hardware type.
6. Review the configuration, accept the terms and conditions, and click Deploy.

Once deployment completes, the model appears in the Deployed list and is available in the Doc Intelligence Node.

Manage deployments

Click any deployment row to open its details across three panels:

Overview — Shows deployment status and configuration: name, language, scaling parameters, and hardware resources.

Deployment history — Lists all past and current deployment versions with status, timestamps, and ownership. Expand an entry to view lifecycle details such as deployment and undeployment events.

Configurations — Manage the deployment lifecycle:

• Undeploy: Deactivates the model. Once undeployed, it is unavailable to workflows.
• Delete: Removes the configuration. Available only after the deployment is undeployed.

Azure Doc Intelligence

Connect external Azure document-processing models for use in the Doc Intelligence Node. You can add and manage multiple Azure connections.

Add a connection

1. Go to Settings → Manage Models → OCR → Azure doc.
2. Click Add a Connection.
3. Enter the required details:
   • Connection Name (must be unique).
   • Base URL.
   • API Key.
4. The system validates the API key and checks for duplicate names.
5. After validation, click Save.

A confirmation message appears, and the connection is added to the list.

Manage connections

From the connection list:

• Enable or disable: Toggle the switch next to a connection.
• Delete: Click the delete icon and confirm to permanently remove the connection.