FieldProvenance

Required fields

fieldmethodextractedAt

Fields

Field	Type	Description
confidence	integer	Confidence in the field value (0-100)
dataSource	object	The third-party source that provided this field value. Enables attribution credits and the data source kill switch
extractedAt	string	When this value was extracted from the source document
field	string	The field name this provenance record applies to
method	object	How the field value was obtained
ocrEngine	string	OCR engine or extraction tool used to obtain this value
originalScript	object	Writing system of the source document. Critical for handwritten wills in non-Latin scripts latin: English, French, German, etc. kanji: Japanese kanji characters. hiragana/katakana: Japanese syllabaries. arabic: Arabic script (also used for Urdu, Farsi). hebrew: Hebrew script. devanagari: Hindi, Marathi, Sanskrit. tamil: Tamil script. chinese_simplified/traditional: Chinese characters. hangul: Korean. cyrillic: Russian, Serbian, etc. thai: Thai script. mixed: multiple scripts in one document. other: any other writing system.
originalText	string	The raw text as extracted before any normalisation or structuring. Preserves exactly what the source document said
sourceDocumentId	string	Reference to the document.json entity this value was extracted from
sourcePageNumber	integer	Page number in the source document where this value was found
sourceRegion	string	Location within the page where this value was found
verifiedAt	string	When this field was last verified
verifiedBy	string	Who verified this field