User Guide Extraction And OCR Extraction Model Settings

Extraction model settings

Configure the artificial intelligence model and token limits used for reading historical documents. Open this screen when you want to change how the program extracts genealogical facts from source documents.

What you see

The upper section contains the structured extraction settings. The extraction model input lets you type or select a model name, and leaving it blank tells the program to use the optical character recognition model. The output token allowance field sets the maximum token budget per document. The Save extraction settings button saves your changes and displays a confirmation message or an error if the save fails.

Common tasks

Change the extraction model

1. Select the extraction model input field.
2. Choose an option from the menu or type a model name.
3. Select Save extraction settings.
4. Read the status message that appears below the button to confirm the save.

Update the token allowance

1. Select the output token allowance per document input field.
2. Enter a new number of tokens.
3. Select Save extraction settings.
4. Read the status message that appears below the button to confirm the save.

Good to know

• Changes take effect only after you select Save extraction settings.
• Leaving the extraction model field blank falls back to the optical character recognition model.