Text recognition and structured extraction

Extract text and structured information from images for search, editing, or downstream systems.

Supported capabilities

Capability	Input	Output
Image text recognition	One image	Plain text
Mathematical formula extraction	Image containing formulas	Extracted and stitched formula image
Invoice and receipt recognition	Invoice, receipt, or sales-slip image	Structured JSON containing recognized fields

The amount, date, tax number, and other fields in a structured receipt result depend on the actual document type and recognition result.