Optical Character Recognition (OCR) Guide: Scanning Text

When you scan a paper document or take a picture of a page with your phone, the output is a raster graphic (pixels). To a computer, a scanned PDF is simply a large image. It has no editable characters, meaning you cannot select text, search for phrases, or copy lines.

Optical Character Recognition (OCR) is the process of scanning these pixels and recognizing letter shapes. Our OCR engine uses advanced neural networks (such as Tesseract WebAssembly) to detect baseline text rows, identify characters, and calculate text coordinate boxes.

Once recognized, the OCR engine generates an invisible vector text layer and aligns it directly on top of the corresponding image pixels. This transforms a flat scanned image into a searchable, selectable, and editable document without changing its original visual layout.