How-to
Make a scanned PDF searchable
A scanned PDF looks like a normal document but is really a picture of a page: Ctrl+F finds nothing, and you cannot select or copy a single word. The PDF OCR tool fixes this without changing what the file is. It reads each page image, recognizes the characters with an on-device OCR engine, and lays the recognized text invisibly on top of that same image, so the PDF still looks and prints exactly as before but now responds to search and text selection. The file never leaves your device.
Step by step
- Open the PDF OCR tool and drop your scanned PDF. Only PDF files are accepted, and the file is capped at 80 MB; a typical multi-page scan is well under that.
- Pick the document's language from the dropdown (English, French, German, Spanish, Portuguese or Italian) and, if you want, edit the output file name. The default keeps the .pdf extension automatically.
- Click Run and wait for every page to be processed. Each page is rendered, run through the OCR engine and given its own invisible text layer before the tool moves to the next one, so a long document takes longer than a single page. Download the result: it opens and prints identically to the original, but Ctrl+F, text selection and copy-paste now work.
Searchable PDF or plain text: which one do you actually need
This tool and the plain OCR tool solve the same underlying problem, recognizing text in a scanned image, but they keep different things. The OCR tool throws away the page image and gives you back a plain .txt file: use it when you want to paste the words into an editor, translate them or search them as text, and you do not care about preserving the document's original look. The PDF OCR tool here keeps the original PDF exactly as it is, page images and all, and only adds a hidden text layer underneath: use it when the document needs to stay a PDF you can still print, email or archive, but you also want to be able to search or copy from it. Neither tool reconstructs the layout as editable text with real fonts and tables; both work from the same recognized text, just packaged differently.
Why the result depends on the scan, and where the limits are
Recognition accuracy tracks the quality of the source image, the same way it does for plain OCR: a clean, straight scan around 200 to 300 DPI recognizes reliably, while a blurry photo, a skewed page or heavy shadows produce more misread words. If a page came out crooked, straighten it with the PDF deskew tool before running OCR so the recognition step reads level text. The tool covers six languages (English, French, German, Spanish, Portuguese, Italian); a document in an unsupported language will not recognize correctly. Processing runs page by page in your browser tab, so a document with dozens of pages takes noticeably longer than a two-page scan, and the file is capped at 80 MB. The output PDF also tends to be similar in size to the original, since the page images themselves are not touched; if you need a smaller file afterward, run it through the PDF compressor.
How the invisible text layer is built
Each PDF page is first rasterized to a canvas image in your browser, the same rendering step the PDF-to-image tool uses. The OCR engine (Tesseract, compiled to WebAssembly) reads that image and returns every recognized word together with its bounding box, its position on the page. Those words are then drawn back onto a copy of the original PDF page at their recognized positions, but at zero opacity, so nothing changes visually: the page image you see and print is the same one that was scanned, sitting on top. A PDF viewer's search and text-selection features only look at that invisible text layer, which is why the file becomes searchable and selectable without looking any different. All of this, rendering, recognition and reassembly, happens inside your browser tab; the PDF is read from local disk and never transmitted anywhere.
The tools used in this guide
- PDF OCR · make it searchableTurn a scanned PDF into a searchable, selectable PDF entirely in your browser, no upload.
- OCR · image/PDF to textExtract text from scanned images or PDFs entirely in your browser, runs offline, no upload.
- Deskew scanned pagesStraighten skewed scanned PDFs or images entirely in your browser, no upload.
- Compress PDFReduce PDF file size by losslessly optimizing its internal structure, without uploading.
Frequently asked questions
Will the searchable PDF still look and print exactly like the original scan?
Yes. The tool never modifies the page image; it only adds an invisible text layer underneath it. On screen and on paper, the output is visually identical to the scan you dropped in. What changes is that a viewer's search and text-selection tools can now find and grab the recognized words, because they read from that hidden layer.
Why does this tool exist alongside the plain OCR tool?
They answer different needs from the same recognition step. The OCR tool gives you a plain text file, which is what you want if you plan to paste, edit or translate the words and do not need to keep the document as a PDF. This tool keeps the file as a PDF with the original page images intact, which is what you want if the document still needs to be printed, archived or emailed as-is, but you also want to be able to search or copy text from it.