What OCR is doing under the hood
OCR (optical character recognition) detects glyph shapes in your image, maps them to characters, and outputs plain text you can paste into email, spreadsheets, or code editors. FormatWiz keeps that pipeline on-device so screenshots of IDs or medical forms are not sent to an API you cannot audit.
Start with the highest-resolution crop you have: rescuing text from a 72dpi fax is harder than from a 300dpi scan.
Getting cleaner text out
Deskew and crop to the paragraph you need before running OCR. Uniform lighting, straight pages, and dark text on light paper all improve confidence scores. If the source is a photo, avoid flash glare on glossy paper.
For multi-column magazines, expect reading order quirks; proofread the output against the original.
Languages and encoding
Modern engines handle many Latin, Cyrillic, and CJK characters, but mixed-language pages may need a second pass with a manual language hint in desktop tools. Export UTF-8 text when you move data into apps that support Unicode end to end.
Sensitive documents
Because recognition runs locally, you can OCR confidential briefings on an air-gapped laptop without network uploads. Still follow your organization's record-handling rules and delete temporary exports when finished.