How to use OCR PDF
- Choose a PDF. The pages are checked at once: pages that already have text are counted and kept as they are, and only the pages that are pictures will be read.
- Pick the language on the pages, English or Urdu, and tap Read. The text reader downloads once (its size is on the button) and stays on your device for next time.
- Watch the pages count up with the time left. Stop at any point: the pages already read are kept, and the saved PDF says which ones have text.
- Save the searchable PDF, or save the text as a .txt file or copy it. Open the PDF and press Ctrl+F (Find on a phone) to search it.
Why use OCR PDF
- Searchable, with the pages as they were
- The PDF you save is your PDF: the same pages, pictures and size, with an invisible text layer over each scanned page. Search it, select and copy its words, and find it by its contents in a file search.
- Pages with text are kept
- A PDF where only some pages are scans is handled page by page: typed pages keep their own text and are not read again, so a 40-page file with 3 scanned pages is done in under a minute on a computer.
- English and Urdu, measured
- On clear single-column English book and letter scans we tested, about 95 to 99 characters in 100 came out right. Printed Urdu in Nastaliq, read by our own model, about 86 in 100. Check names and numbers afterwards.
- Private, and honest about time
- Every page is read in your browser; the text reader downloads once and then works offline. A scanned page takes about 13 seconds on a computer and longer on a phone, and the bar shows the real time left.
About this tool
OCR (optical character recognition) turns the picture of a page into text. A PDF made by a scanner, a copier or a "save as PDF" of photos holds only pictures, which is why Ctrl+F finds nothing in it and the text cannot be selected. Most online PDF OCR services upload the file to a server; this searchable PDF converter does the work in your browser instead. It draws each scanned page at about 200 dots per inch, finds the lines of print, reads them, and writes every word back into the PDF at its exact place in an invisible font. The page still looks like the scan, but a PDF reader, a phone's Files app or a desktop search now sees the words. Pages that already carry text, such as a typed report with one scanned signature page, are left exactly as they are. If a PDF's own text is wrong (copied text comes out as nonsense, which happens with some fax and copier software), switch on "Read every page as a scan" to read the pictures instead.
What it reads well and what it does not: clear printed English in one column, such as books, letters, statements, forms and receipts, reads with about 95 to 99 characters in 100 right on the scans we measured. Printed Urdu in Nastaliq or Naskh reads with about 86 in 100 using our own Urdu model, the one behind the Urdu OCR page. Newspapers and other pages with several columns come out with the columns mixed into each other, handwriting is not read, and Arabic is not offered here because it did not meet our bar on scanned books (for a single Arabic page, save it as a picture and use the Arabic OCR page). A photo of a page taken with a phone, at an angle or in shadow, reads better through Scan to PDF, which finds and straightens the page first and can then make the PDF searchable too.
Frequently asked questions
How do I make a scanned PDF searchable?
Choose the PDF, pick English or Urdu, and tap Read. Each scanned page is read on your device and the words are written invisibly over it. Save the PDF, open it in any reader and press Ctrl+F (or Find on a phone) to search it.
Is my PDF uploaded?
No. The PDF is opened, read and saved in your browser; nothing is sent to a server. The text reader is downloaded once, to your device, and the site never sees the file or its text.
How long does it take?
About 13 seconds a scanned page on a computer, and longer on a phone, so a 20-page scan is about 4 minutes on a laptop. Pages that already have text take no time. The bar shows the real pages done and the time left, and you can stop at any point and keep the pages read so far.
What if the PDF already has text?
Then there is nothing to add: the page tells you so and offers the text as a .txt file. If that text is wrong (some scanners add a poor text layer), switch on "Read every page as a scan" and the pictures are read instead.
Does it read Arabic PDFs?
Not yet. On the scanned Arabic books we measured, only about 66 characters in 100 came out right, below the bar we set, so Arabic is not offered on this page. For a single Arabic page, save it as a picture and use the Arabic OCR page, which reads Arabic text in images.
Why does my newspaper or two-column page come out jumbled?
The reader takes each line across the whole page, so two columns are read as one wide line and their words interleave. Single-column pages (books, letters, forms) read well; for a newspaper, crop each column to its own picture and use Image to Text.
Does it work on an iPhone or Android phone?
Yes. Pick the PDF from Files or Downloads in Safari or Chrome, read it, and save the searchable PDF to Files or share it. Nothing is installed. A long scan takes longer on a phone than on a computer, so keep the page open while it reads.
Can I copy the text instead of saving a PDF?
Yes. After the pages are read, copy the text or save it as a .txt file, page by page. To turn a PDF into an editable Word document with its layout, use PDF to Word instead.
Built and maintained by Saad Naeem
AI engineer, Doha · Updated