Label Version

The 'Text' property specifies the exact string of the label as it appears in the document. This is the value that label-aware extractors and classifiers will search for when identifying data elements.

• Enter the label text exactly as it appears, including case, punctuation, and line breaks if applicable.
• For multi-line labels, use line breaks to match the document layout.

Example:

• "Invoice Number"
• "Inv #"
• "Total\nAmount" (for a vertical or wrapped label)

Use multiple Label Version entries to capture all variations.

Can be one of the following values:

The 'Layout' property defines how the label text is arranged in the document, such as horizontal, vertical, or multi-line. This helps label-aware extractors match the label even when its orientation or formatting changes.

• Choose the layout that matches the appearance of the label in your documents.
• Common options include horizontal (single line), vertical (stacked), or custom arrangements.

Example:

• Horizontal: "Invoice Number"
• Vertical: "Invoice\nNumber"

Accurate layout selection improves extraction reliability, especially in complex or variable document formats.

The 'Padding' property allows you to specify extra space around the label's bounding box, in logical units (such as mm, in, pt). This can help account for slight misalignments, font variations, or to ensure the label region fully encompasses the text.

• Adjust padding to fine-tune label detection, especially when labels are close to other content.
• Use logical units (e.g., "2mm", "0.1in") for precise control.

Example:

• "2mm" adds 2 millimeters of padding on all sides.
• "0.1in, 0.2in" sets left/right and top/bottom padding independently.

Proper padding can improve extraction accuracy in documents with tight layouts or variable spacing.