- Document Understanding para Automation Cloud y Test Cloud
- Document Understanding para Automation Cloud y Test Cloud Public Sector
- Document Understanding para Automation Cloud y Test Cloud Dedicado
- Paquetes ML y puntos finales públicos
Notas de la versión de Document Understanding
26 de junio de 2026
Introducing Field Groups and Monetary Quantity
Field Groups and Monetary Quantity are now generally available in Validation Station, activities, and APIs. For more details, refer to Taxonomy Manager.
This enhancement aligns the Document Understanding data type model with the IXP Unstructured and Complex Documents.
Document Understanding API v2
Document Understanding API v2 is now generally available. API v2 adds taxonomy-driven extraction and validation, business rule evaluation, expanded non-Latin language support, and Intelligent Xtraction and Processing (IXP) FieldGroup support in the extraction result.
Novedades
- Discovery returns the associated taxonomy in the response so you can use it to guide extraction and validation.
- Extraction and Validation accept taxonomy as an input parameter. This enables partial extraction, data type overrides, and additional validation customization. For details, see Use the Validation Service.
- La API evalúa las reglas empresariales definidas en la taxonomía durante la extracción y la validación para garantizar que los datos extraídos cumplan con los requisitos configurados, como los campos obligatorios y los valores permitidos.
- Los proyectos predefinidos ahora admiten la extracción basada en etiquetas utilizando una sola etiqueta llamada producción.
- IXP extraction result is now returned as FieldGroups instead of Tables. With FieldGroups, each field preserves its actual data type. For details, see Taxonomy Manager.
- A new predefined project offers enhanced support for non-Latin character sets.
Document Understanding API v2 breaking changes
The following changes require updates to your automations when migrating from API v1 to v2:
- Response bodies on all routes now use PascalCase property names instead of camelCase. Any case-sensitive integrations, such as those written in Python, JavaScript, or others, require updating to match the new casing.
- Tag-based routes now use
/projects/{projectId}/tags/{tag}/.... Requests using the previous route structure return a 400 bad request error in v2. - The
fieldsproperty has been removed from Discovery responses; integrations must usetaxonomyinstead. - Las rutas de descubrimiento ahora devuelven
tagsen lugar detag. - IXP extraction results are now returned as FieldGroups instead of Tables, which requires updates if your automation treats extraction results as tables to handle the new FieldGroup structure and data types.
- Entries that previously returned
FieldType == Tablenow returnFieldType == FieldGroup. Parsers that handle onlyTableresults will not process this data. For detailed migration steps and examples, check Migrating integrations from Document Understanding API v1 to v2.
8 de junio de 2026
Corrección de errores
Las páginas en blanco provocaban que los resaltados de extracción aparecieran en las páginas incorrectas en Action Center. Los resultados de la extracción no se vieron afectados.