Convert Scanned PDF to Word
Choose a scanned PDF, and a neural network reads the text on every page. You get a Word document, a TXT file and the same PDF with text you can search.
Files are processed on your device and never uploaded to a server.
- Choose PDF
- Choose the language
- Download the text
Recent results
The first time, the browser downloads a 20 MB neural network. After that it works offline.
Recognizing text…
The graphics card is not available for the neural network on this device, so the processor does the work. It takes longer, but the result is the same.
Done! Text recognized
What's next?
More things to do with your document
How to use
- Click "Choose PDF" or drop the file onto the page. Scans and PDFs made from photos both work. On a phone, the file can be up to 100 MB.
- Pick the document language (English is preselected) and click "Recognize text". The first time, your browser downloads text recognition: 4 to 20 MB depending on the language, about 20 MB for English. Next time it loads from the browser's storage.
- Click "Download DOCX". The PDF with text and the TXT file are right next to it, under "Download PDF" and "Download TXT".
What you get from the scan
A Word document with the text of every page. Lines of one sentence are joined into paragraphs, and columns come out in reading order. Each PDF page starts on a new page in Word. Tables and pictures don't go into Word, only the text.
A PDF with text: the scan looks the same, but invisible text now lies under each page image. Ctrl+F finds words in it, and you can select and copy them. If some pages already had text, they aren't recognized again: that text comes straight from the file.
How to convert a scanned PDF to Word with fewer errors
A sharp scan means fewer typos. Dark or tilted pages are read worse, and handwriting isn't recognized at all.
Set the document language before you start. Each alphabet has its own recognition model, so with the wrong language the text comes out garbled.