The short answer
Choose the language of the document, drop the scanned PDF on the tool below and wait while it reads the pages one by one. When it is done, download the searchable PDF, the TXT file or both in one ZIP. The searchable PDF keeps every original page untouched and only adds the invisible text, so it looks identical and grows only slightly in size.
If the document mixes two languages, for example a German contract with English terms, pick a second language as well. If you only need some pages, type them in the Pages field, for example 1-3, 5 or 8-.
Try it here: OCR PDF
Open the full tool pageWhy a scanned PDF is not searchable
A PDF saved from a word processor stores real text: every letter, its font and its position. A scanner, a copier or a phone scanning app works differently. It takes a picture of the paper and wraps that picture in a PDF. To you the page looks like text, but to the computer it is a grid of coloured dots.
That is why the search box finds nothing, the cursor selects the whole page as an image, and pasting gives you nothing useful. Screen readers cannot read it aloud either, and file search on your computer or in a document archive cannot find the document by its content. Some scanning apps run OCR automatically, which is why some scans are searchable and others are not.
A quick test: open the PDF, press Ctrl+F (Cmd+F on a Mac) and search for a word you can clearly see on the page. If nothing is found, the page has no text layer and needs OCR.
How the invisible text layer works
OCR finds lines and words on the page image and recognizes each character. The tool then writes every recognized word into the PDF at the exact place where it appears in the scan, in an invisible text mode that PDF readers know: the text is there for searching and selecting, but it is not drawn, so the scan underneath stays visible as it was.
The original pages are copied as they are. The scan is not converted again, not compressed again and not reduced in quality, so the result looks identical to your file. Each page only gains a few kilobytes of text. When you search, the reader highlights the word on the scan, and when you select a line and copy it, you get the recognized text.
The settings, one by one
- Language of the text: the language the document is written in. Choosing the right one matters most, because it tells the OCR which letters and words to expect. Only the data for that language is downloaded, once.
- Second language: optional. Use it for mixed documents. Both languages are recognized together, which is a little slower than one.
- Pages: leave it empty for all pages, or enter ranges such as 1-3, 5 or 8- (from page 8 to the end). The searchable PDF still contains every page; only the chosen ones get a text layer.
- Result: searchable PDF and text (both, as a ZIP), searchable PDF only, or text only. Changing it later does not read the pages again, the recognized text is reused.
- Stop: long documents can be stopped at any time with the button next to the file. Pages already read are kept, so the next run continues where it stopped.
Tips for accurate text
- Scan at 300 DPI if you can. Pages are rendered at about 300 DPI before reading, but a sharper scan always gives better letters to work with.
- Keep pages straight and evenly lit. Slight skew is fine, strongly rotated or curved pages, such as photos of a thick book, read worse.
- Choose the exact language. With English selected, letters like ç, ğ, ş or ß can come out as similar looking Latin letters.
- Check names, amounts and dates before you rely on them. OCR is very good on clean print, but a 1 can turn into an l and a 0 into an O on poor copies.
- Handwriting, stamps and text on photos or busy backgrounds are hard for any OCR and may be skipped or misread.
Which languages work, and how well
Recognition covers 20 languages. The text in the TXT file is always plain Unicode, so every alphabet is kept. In the searchable PDF the invisible text is stored character by character, so letters such as ç, ğ, ı, ş, é, ñ and ß are searchable and copy correctly, which is not the case with many simple PDF tools that only support basic Western letters.
| Script | Search and copy in the PDF | Notes |
|---|---|---|
| Latin (English, Turkish, German, French, Spanish...) | Yes | Accents and special letters kept |
| Cyrillic and Greek | Yes | Works in all common readers |
| Chinese and Japanese | Yes, with limits | Words may be split by spaces when searching |
| Arabic, Hindi, Thai, Korean | Yes, with limits | Some readers order or split words differently |
Pages that already have text, and your privacy
Many PDFs mix scanned pages with normal ones, for example a typed cover letter followed by scanned receipts. Before reading a page, the tool checks whether it already contains real text. If it does, the page is left exactly as it is and its existing text goes into the TXT file, so you never get a second, duplicated text layer. A PDF that was already processed with OCR is detected the same way.
Everything happens on your device. The PDF is opened, rendered and read in your browser, and nothing is uploaded, which matters for contracts, medical letters and bank statements. After the first use the language data stays in your browser, so the tool also works offline. Password protected PDFs cannot be opened; remove the password first.