OCR, optical character recognition, reads the letters in a picture of a page and turns them into text a computer can search. A scanned PDF is just pictures, so you cannot search it, select a sentence or copy a number from it. After OCR you can. This tool does it in your browser, with no page limit, and your document is never uploaded.
Make a scanned PDF searchable
The result looks exactly like your scan, because the original pages are kept. Under the picture, an invisible layer of recognized words sits where each word appears, so search highlights the right spot and selecting text picks up the words you see. To check whether a PDF already has text, try selecting a line: if you cannot, it needs OCR.
Choose OCR languages
Recognition needs to know which alphabet and words to expect, so pick every language that is printed in the document, such as English and Spanish for a bilingual form. Over 100 languages are available, including Chinese, Japanese, Korean, Arabic, Hebrew and Hindi. Each language is downloaded once, usually a few megabytes, and then kept by your browser; the size is shown before you start. Picking languages that are not in the document makes recognition slower and less accurate.
What affects OCR accuracy
On a clean, straight scan of printed text at 300 dpi, nearly every word is read correctly. Accuracy drops with blurred phone photos, low resolution, skewed or curled pages, faint print, colored backgrounds, unusual fonts and tables with tight lines. Handwriting is mostly not recognized. Raising the resolution in the advanced settings helps small print, at the cost of time. Always check names and numbers you rely on.
How it works
Each page is drawn as an image in your browser and read by Tesseract, an open source OCR engine, running on your device. The words it finds and their positions are written into a copy of your PDF as an invisible text layer, and collected into the text file. A laptop handles a page in a few seconds; a phone is slower. Long documents take longer but have no limit, and the estimate before you start tells you what to expect.
What people use it for, and when it will not work
Scanned contracts and letters, old books and theses, receipts and invoices for bookkeeping, school handouts, records from an archive, and anything you want to search on your computer later.
- Layout is not rebuilt. The searchable PDF keeps the original look; the text file is plain text. To edit the document, use PDF to Word afterwards.
- Handwriting is mostly not recognized.
- PDFs that already have text do not need OCR; to get their text, use PDF to Text.
- On a phone, keep the screen on and the tab open during long jobs, or the browser pauses them.