SDK Samples

Extracting data from images using ICR

Use the Nutrient SDK's ICR engine to extract structured document data from images with local models. Best on printed pages and clean print handwriting in structured layouts; cursive narrative content is outside its current strength.

Source Document

What ICR is good at

Printed text on multi-language pages, and clean hand-printed text in known form layouts — labels and short field values like names, addresses, and phone numbers. Cursive, narrative pages, and decorative calligraphy are outside the engine's current strength.