What PDF OCR does
A scanned PDF is a set of pictures. You can read it, but your computer cannot: searching finds nothing and text cannot be copied. OCR — optical character recognition — reads those pictures and turns them into real text. This tool gives you the recognised text and a searchable copy of your PDF, in which an invisible text layer sits exactly behind the scanned image. English and Hindi are supported, and recognition runs in your browser.
How to OCR a PDF
- Add your scanned PDF.
- Choose the document language: English, Hindi or English + Hindi.
- Choose the output: text plus searchable PDF, or text only.
- Press start. Pages are recognised one by one and a progress bar shows where the tool is.
- Copy the text, download it as a file, or download the searchable PDF.
The first run downloads the recognition engine and language data (about 9 MB). Your browser caches it, so later runs start faster.
What you get
| Output | Use it for |
|---|---|
| Recognised text | Pasting into an e-mail or document, or editing |
| Searchable PDF | Archiving — the file looks like the scan, but you can search, select and copy |
| Confidence score | A quick indication of how reliable the reading is |
What affects accuracy
| Factor | Good | Poor |
|---|---|---|
| Resolution | 200–300 dpi scans | Low-resolution phone screenshots |
| Alignment | Straight pages | Tilted or curved pages |
| Contrast | Dark print on white paper | Faded print, coloured or patterned backgrounds |
| Content | Printed text | Handwriting, stamps, decorative fonts |
Clean printed pages usually read very well. Handwriting is not reliably recognised. Always proofread names, numbers and dates before relying on them — OCR can confuse similar shapes such as 0 and O, or 1 and l.
Use cases
- Making scanned contracts, letters and records searchable.
- Copying a paragraph from a printed page without retyping it.
- Preparing a scan for PDF to Word or PDF to Excel, which need real text.
- Building a searchable archive of paper documents.
Tips for better results
- Scan at 300 dpi if you can, in greyscale or black and white.
- Straighten sideways pages first with Rotate PDF.
- Crop away dark scanner borders; they can be misread as characters.
- Choose the correct language. Use English + Hindi only for documents that really mix both, because the combined mode is slower.
- For long documents, split the file and process it in parts — OCR is demanding and phones have limited memory.
Good to know
OCR takes time: expect a few seconds per page on a computer and longer on a phone. Keep the tab open while it works. If your PDF already contains selectable text, you do not need OCR — use Extract Text from PDF instead.
Privacy
Recognition runs in your browser on your own device. The PDF is not uploaded to our server.