OCR PDF
Read the text in a scan and make the PDF searchable.
Drop your file here
or choose from this device — a PDF or a photograph
What this actually does
Recognition runs on your device, so the first run downloads a language model of about 15 MB and a long document takes a while. Accuracy depends on the scan — clean 300 dpi text is near perfect, a phone photograph of a curled page is not. The searchable PDF carries a real text layer in all three languages, Hindi and Gujarati included, placed word by word over the ink so that selecting a line selects what you dragged across.