← Blog

How to OCR a Scanned Document Into a Searchable PDF on Windows

A scanned document is just a picture of text. You can see the words, but your computer can’t — so you can’t search it, copy from it, or index it. OCR fixes that by reading the words and adding an invisible, searchable text layer behind the image.

Most OCR problems with scans are really scanning problems. So this guide starts with how to scan for good OCR, then covers adding the text layer with free Windows tools, and finishes with how to check the result.

When you need this

  • Scanned contracts, forms, or receipts you need to search later.
  • Old documents you photographed or scanned to a folder of images.
  • Any PDF that’s really just images with no selectable text.

Scan for OCR, not just for looks

If you are scanning fresh, a few settings make a big difference:

  • Resolution: 300 dpi is the usual recommendation for text. Lower resolutions save space but small print gets hard to read; much higher resolutions mostly make files bigger.
  • Color mode: grayscale is a good default for text documents. Use color only when color carries meaning (stamps, highlighted passages, charts).
  • Straight pages: line the page up against the scanner edge. Tilted text is harder to recognize, although some tools can straighten (deskew) pages automatically.
  • Clean glass: dust and smudges become specks that OCR may try to read as punctuation.

Phone photos of documents can work too, but shoot straight down, in even light, without your shadow across the page. A slight angle makes lines curve, which hurts accuracy at the edges.

Free ways to add OCR on Windows

NAPS2: scan and OCR in one app

  1. Install NAPS2 (free and open source).
  2. Either scan directly from your scanner inside NAPS2, or import existing scan images or PDFs.
  3. Rotate, reorder, or delete pages in the thumbnail view.
  4. Enable OCR and choose the language of the document.
  5. Save as PDF. The output has a text layer behind each page.

OCRmyPDF: for scans that are already PDFs

OCRmyPDF is a free command-line tool that takes an existing PDF and writes a new one with a text layer. It has options to deskew crooked pages and to rotate pages that were scanned sideways, and you can set the OCR language. It’s a good fit when you have many scanned PDFs, because you can run it on a whole folder with a short script.

When the document is only on screen

Sometimes a scan lives inside a viewer you can only look at, such as a document portal or an app that displays pages but won’t export them. Then the job becomes: display each page, capture it cleanly with the margins cropped out, OCR it, and bind the pages into one PDF. Doing that by hand for every page is the slow part.

CapDrop does this whole chain in one pass on Windows: pick a region and a page key, set an interval, and walk away — it captures each page, auto-crops, runs OCR locally, and binds everything into one searchable PDF. 100% local, for content you own, 7-day free trial. Try a real searchable sample — Ctrl-F it yourself.

How to confirm the OCR worked

  1. Open the finished PDF, press Ctrl-F, and search a word you can see. It should highlight in place.
  2. Select a few lines with the cursor. The highlight should follow the words, not jump around.
  3. Copy a paragraph into Notepad and read it. This is the fastest way to spot recognition errors.
  4. Check numbers carefully. In contracts and receipts, a 5 read as an S or a 0 read as an O can make a search fail even though the text “looks” fine.

Common pitfalls with scans

  • Two-page spreads. A book scanned open flat puts two pages in one image. OCR handles it, but text near the binding is often curved. Scan one page at a time if you can.
  • Faint or faded originals. Increasing contrast in the scan settings often helps more than any OCR option.
  • Tables and forms. Text in boxes is usually recognized, but the table structure is not preserved in copied text. Search still works.
  • Handwriting. Standard OCR engines are built for printed text. Handwritten notes and signatures generally won’t become searchable.
  • Overwriting the original. Always save the OCR’d version as a new file and keep the original scan until you’ve checked the result.

A note on privacy

Scanned documents are often the most sensitive files you own: contracts, IDs, medical forms, tax records. Free online OCR sites work by uploading the file to their servers. The desktop tools above all run on your own PC, and CapDrop’s OCR is local too — your documents are never uploaded.

FAQ

Will OCR change how my scan looks? It shouldn’t. A searchable PDF keeps the original image and adds text behind it. Some tools can optionally clean or compress images, so check the settings if you need an exact copy.

Can I OCR documents in other languages? Yes, with the right language selected. The free tools above support many languages through the Tesseract engine, but you may need to install the language data first.

Why does search miss a word I can clearly see? Usually a recognition error in that word. Copy the surrounding text into Notepad to see what the OCR actually produced.


Get CapDrop →