We released the following capabilities as part of the Box Extract APIs:
- Citations that refer to text segments or verbiage within the document being extracted and are used as the basis for extracting data for specific metadata fields
- Bounding boxes that visually indicate and highlight the citations listed above.
These new APIs are available for both the Box Standard and Enhanced Extract Agents. There may be multiple citations or bounding boxes for a single metadata field for a document that us being extracted. For example, the Extract Agent may need to combine a contract start date and contract duration to figure out the contract end date. In this situation. the citations and bounding boxes would pertain to both the contract start date and a contract duration mentioned in the document.
To learn more about this release, please see the Structured Endpoint Documentation.