OCR (Optical Character Recognition)

To a computer, a photo of your CV is just coloured dots — there are no words in it to copy or search. OCR software finds the areas that contain text, recognises each character and rebuilds the lines and words. Good OCR engines also straighten tilted pages, cope with different fonts and support several languages, since accents and special letters — such as Turkish ç, ğ, ı and ş — need the right language data to be read correctly.

OCR is the first step whenever a CV arrives as an image; resume parsing then works out what the text means.

Many employers' systems don't run OCR at all, or run it with mixed results — so a CV that exists only as a scan or a picture may arrive as an empty or garbled record. Typical OCR mistakes are confusing similar characters (0 and O, 1 and l), splitting columns in the wrong place and missing light text on dark backgrounds. When you send a CV to an employer, send a text-based file; keep scans for cases where you have nothing else.

Twin CV's CV import runs OCR on scanned PDFs and photos (JPG, PNG, WebP) in English and Turkish. It reads the first three pages of a scan, turns sideways pages upright and makes a second pass for light text on dark panels. The result fills your CV sections, and you can then download it as a clean, text-based PDF. If nothing can be read, it asks for a sharper photo — or paste the text into the chat.