mypdf.space

Guide

Make a scanned PDF searchable

A filing cabinet that has been scanned is a filing cabinet you cannot search. The documents are safely digital and completely opaque: no Ctrl+F, no copy and paste, no way to find the one invoice you need without opening each file and reading it.

The fix is to add a text layer. OCR reads the page images, works out what the characters are, and writes the result into the file as invisible text positioned exactly over the matching words. The pages look identical — but now the file is searchable, and so is your whole folder once your operating system indexes it.

Step by step

  1. 1Open OCR PDF and drop in the scanned document.
  2. 2Select the language used in the document.
  3. 3Start recognition and let it work through the pages on your device.
  4. 4Download the result — same pages, now with a hidden text layer.
  5. 5Open the file and press Ctrl+F (Cmd+F on a Mac) to search for a word you can see on the page.
  6. 6If the file grew, run it through Compress PDF in Keep text mode to bring the size back down.

What "searchable" actually means

The recognised text is placed in the PDF with a rendering mode that makes it invisible, positioned to match the words in the underlying image. Nothing about the appearance changes — print the file and it is identical to before.

What changes is everything a computer can do with it. Search works. Copy and paste works. Screen readers can read the document aloud. Desktop search indexes the contents, so the file turns up when you search your machine for a phrase rather than a filename. And conversion tools finally have something to export, which is why this same step is the prerequisite for turning a scan into Word.

Scan quality decides the outcome

Three things dominate accuracy, and all of them are set before OCR runs. Resolution: aim for 300 DPI. Below 200, characters blur into each other and the engine starts guessing. Above 400, you mostly buy a longer wait.

Straightness: skew is the underestimated one. A page fed slightly crooked through a scanner produces text on a gentle diagonal, and recognition rates drop noticeably. If your scanner offers deskew, use it.

Contrast: clean black on white beats a grey third-generation photocopy by a wide margin. If the source is faint, rescanning in black-and-white mode usually helps more than any adjustment afterwards.

Working through a backlog

Browser OCR is not fast, because it is doing real work on your hardware rather than in a datacentre. For one document that is fine. For two hundred, be strategic: run them on a laptop rather than a phone, do them in batches, and start with the files you are most likely to need to find again.

It is also worth doing the OCR before you compress. Recognition works from the page images, so giving it the higher-quality original produces better text — then compress the finished, searchable file.

Tools used here

Frequently asked questions

  • Will OCR change how my document looks?
    No. The original page images are preserved and the recognised text is added as an invisible layer on top. Printed or viewed, the document is unchanged.
  • How long does it take?
    Roughly a few seconds per page on a modern laptop, longer for dense pages or on a phone. The first run also downloads the language data, which is a one-off.
  • Which languages are supported?
    Choose the document language before running recognition. Picking the wrong one produces markedly worse results, because the engine uses the language to resolve ambiguous characters.
  • Can I search a scanned PDF without OCR?
    No. Search operates on text, and a scan contains only images of text. Some viewers offer live OCR while you read, but that does not change the file — to make the document searchable everywhere, including in desktop search, the text layer has to be written into it.

Other guides