コンテンツにスキップ
PDFLeanドキュメント
Proにアップグレード

Make a Scanned PDF Searchable

このコンテンツはまだ日本語訳がありません。

A scanned PDF is a picture of a page. You cannot search it, select its text, or copy from it, because the words are pixels, not characters. Making it searchable means one thing done well: add a text layer over the image, recognize the words on the device, and then verify the result — because recognition is probabilistic and the verification is the difference between a searchable PDF and a file that only claims to be one.

The Job, in Order

Section titled “The Job, in Order”

1. Clean the capture first. enhance a scanned document — in your browser, free — straightens crooked pages and evens out contrast. It is cheap to run and it is the single biggest factor in how well the next step reads the page: recognition follows the quality of what it is given.

2. Recognize the text on the device. OCR a PDF — on-device recognition, free, no upload — reads each page and adds the text layer in place. Nothing is transmitted; the file becomes searchable on your own machine.

3. Verify. Search for a word that must be in the document. Select a line of text. Copy a sentence and compare it with the page. If the search finds nothing or the copied text reads wrong, the problem is the source quality — re-capture or re-clean rather than running another pass on the same pixels.

After the Text Layer

Section titled “After the Text Layer”

Two follow-up decisions, both covered elsewhere in this section:

  • Size. A text layer adds bytes on top of the images, so scans often get heavier once they are searchable. Compress after recognizing, not before — the scan-compression guide walks the trade without blurring the text you just recovered, and compress an image is the lever behind it.
  • Long-term storage. If the file is being kept for years, convert a PDF to PDF/A — in your browser, free — stores it in the ISO preservation format.

Frequently Asked Questions

Section titled “Frequently Asked Questions”

Does this work on handwriting? No. Printed and typed text is what recognition reliably recovers; handwritten pages stay pictures. Any tool that promises otherwise on a full page of handwriting is overpromising.

What about languages? Recognition works on the languages the on-device engine supports, and a page mixing scripts is harder than a single-language page. If a page comes back poorly, check that the original is clean and legible before blaming the step.

Why can I see the text but not search it? Because you are looking at the image, not a text layer — which is exactly the state this guide exists to end.

A scanned PDF becomes searchable when the text layer is real, and real means verified: search it, select it, copy it. Clean the page, recognize on the device, confirm with a search — and the scan stops being a picture and starts being a document, without ever being uploaded.

Do It Directly

Section titled “Do It Directly”

Ready to make the file in front of you searchable? enhance a scanned document the capture first — in your browser, free — then OCR a PDF with on-device recognition, no upload. Search a word that must be there, and copy a line to check. Heavy result? compress an image the source images after recognizing; keeping it for years? convert a PDF to PDF/A.

Section titled “Related reading”