How to Make a Scanned PDF Searchable with OCR

Why you can't search a scanned PDF

A PDF exported from Word or Google Docs carries its text as actual characters, so search, selection, and copy & paste all work. A PDF produced by a scanner or a phone camera carries only images. To your PDF reader, the word “Invoice” on a scanned page is just a pattern of dark pixels — there is nothing to search for.

What OCR actually does

OCR — Optical Character Recognition — analyzes each page image, finds the shapes that form letters and words, and works out what they say and exactly where they sit on the page. A good OCR tool then writes that text back into the PDF as an invisible layer positioned precisely under the original image.

The result looks identical to your scan, but behaves like a born-digital document:

• Ctrl+F / Cmd+F finds words anywhere in the file.
• You can select passages and copy them into email, Word, or a spreadsheet.
• Screen readers and other accessibility tools can finally read the document aloud.

Making a scan searchable, step by step

1. Open the OCR PDF tool and drop in your scanned PDF.
2. Pick the document's language — English, Indonesian, or both for mixed documents. The first run downloads the recognition engine (about 6 MB) into your browser; after that it is cached.
3. Click Recognize Text and watch the per-page progress. On a typical laptop, expect a few seconds per page.
4. The searchable PDF downloads automatically. You can also grab the recognized text as a plain .txt file for editing or archiving.

Everything above happens on your device. The OCR engine (Tesseract) runs as WebAssembly inside the browser, which means scans of contracts, ID cards, and medical records are never uploaded anywhere.

Getting the best recognition accuracy

OCR quality is mostly decided before the software ever runs — at scan time:

• Resolution: scan at 200–300 DPI. Below that, letters blur together.
• Alignment: keep pages straight. Heavily skewed text is much harder to read.
• Lighting: when photographing documents with a phone, use even light and avoid shadows across the text.
• Contrast: black text on white paper recognizes far better than faded print or colored backgrounds.

Even with a perfect scan, expect the occasional misread character in stylized headings or stamps — OCR is excellent, not infallible. For anything legally sensitive, proofread the extracted text.

After OCR: tidy up the document

Once your scan is searchable, the other browser tools apply as usual: merge multiple recognized scans into one archive, split out the pages you need, or compress the result if the original scan was heavy.