Which language does it recognise?

Whichever language packs Windows has installed. If there is none, the app opens the right Windows screen for you to add one — recognition quality follows the pack, not the app.

Does the page look different afterwards?

No. The image of your scan is untouched: the text layer is invisible and sits underneath it. What changes is that selecting, copying and Ctrl+F start working.

Why does Ctrl+F still miss one specific word?

Recognition probably got that word wrong — a stamp on top, creased paper, low contrast. Try searching a neighbouring phrase. If the whole document is poor, a fresh 300 DPI scan usually helps more than changing software.

Will this make Windows Search find my PDFs?

It helps: after OCR the file contains real text, so indexers that read PDFs can see it. Just remember Windows Search still needs that folder in its index.

How long does a big scan take?

Minutes for a few hundred pages, on your own CPU. It runs in batch with a progress bar and a Cancel that frees the desk right away — cancelling discards whatever that run had produced.

How to make a scanned PDF searchable, 100% on your own PC

You open the scanned contract, press Ctrl+F, type the clause number and nothing comes back. The text is right there, in front of you, on the screen. And the computer still swears it doesn’t exist.

It isn’t a bug and it isn’t your fault: a scanned PDF is a photograph of paper. To the program showing the file, that page has no words — it has pixels that happen to look like words to a human eye. This guide shows how to turn those pixels into real text without the document ever leaving your computer, and what to expect from the result.

Why Ctrl+F fails on a scan#

A PDF is a container. It can hold two very different things:

Real text (a “text layer”), when the file was born digital — exported from Word, generated by an invoicing system, saved from a browser.
A picture of the page, when the file was born from a scanner, an office copier or a phone camera.

In the second case there is nothing to search: Windows Search and your reader’s Ctrl+F look for text, and all that exists there is an image. No indexing setting fixes that, because it isn’t a settings problem — it is the difference between text and a photograph of text.

The bridge between the two worlds is called OCR (optical character recognition): software looks at the image, recognises the shapes as letters and produces text from them.

Two different things people call OCR#

Worth separating, because the confusion costs time:

OCR for search — the program reads the image and keeps the words in its own index, so it can find the file later. The PDF itself stays exactly as it was. That is what the Elegant File Explorer Finder does when you switch deep search on.
OCR written into the file — the program creates a new copy of the PDF with an invisible text layer underneath the image. From then on, Ctrl+F works inside that document, in any reader, on any PC, forever.

This guide is about the second one: it is what solves “I need to send this client a searchable PDF”.

Doing it without uploading the document#

Most online OCR services work well — and all of them ask for the same thing: a copy of your document on their server. For a leaflet, fine. For a scanned medical report, contract or payslip, that is handing over exactly what you shouldn’t.

Windows 10 and 11 ship with a recognition engine built in, the same one other system apps use, and it works with no internet, using the language packs installed on the machine. That engine is what Elegant Paper uses: nothing is sent anywhere, and the app has no HTTP client. The only network connection it makes is the licence and trial check against the Microsoft Store, which never sees your files — and is fail-open if the Store is down.

Step by step (under a minute)#

Put the files on the desk. Drag the scanned PDFs onto the Elegant Paper window, or use Add…. Every file becomes a card with its cover; to work page by page, click Expand.
Check what goes in. Pages lying sideways? Rotate. A blank sheet in the middle of the pile? Delete page. OCR reads whatever is on the desk, so it pays to tidy first.
Click “Make searchable”. The button sits on the top bar (and in the More menu). The app starts reading the pages and shows progress: “Reading the text of the pages (OCR)…”.
Wait — or cancel. Large documents take minutes. Cancel is always within reach and frees the desk right away; whatever had been produced is discarded.
Take the new file. The output lands next to the original, with a suffix in the name. The original stays exactly as it was, and a 12-second Undo toast sends the created file to the Recycle Bin if you change your mind.

That’s it: open the new copy, press Ctrl+F and look for that clause. Now it exists.

The language matters (this is where most people go wrong)#

Recognition uses the language packs installed in Windows. If your document is in Spanish and the machine only has English, quality collapses: accents turn into noise and whole words come out wrong.

Elegant Paper’s settings have an OCR language list showing what is installed. If no language with recognition is available, the app opens the right Windows screen for you to add one — under Settings → Time & language → Language & region, when adding a language, tick the optical character recognition feature.

What to expect from the quality#

No OCR is perfect, and be suspicious of anyone promising 100%. In practice:

A clean 300 DPI scan, black text on white: excellent. Ctrl+F finds almost everything.
A slightly crooked phone photo with a shadow: it works, but it makes more mistakes. The app’s Straighten fixes the page geometry before reading — maths, not AI, so no pixel is invented. Even so, retaking the photo with the whole page in frame and decent light beats any correction.
Handwriting: don’t count on it. The engine is built for printed text.
Stamps, signatures, tables with hairline rules: the surrounding text comes out fine; what sits under a stamp, not always.
Old faxes and copies of copies: uneven. If the document matters, a fresh scan fixes more than any software will.

The text layer goes underneath the image: the file still looks like the scan you know, and nobody sees odd letters printed over the document. If recognition gets a word wrong, what changes is the search — not the PDF’s appearance.

Can I do it in batch?#

Yes, and that is the common case: the office copier spits out dozens of files a week. Put them all on the desk and hit Make searchable once. The app processes them in a queue, with progress, and the desk stays locked only until it finishes — cancelling frees it immediately.

Two honest notes about batches:

Time scales with pages, not files. A 600-page bundle takes far longer than 60 one-page receipts.
OCR does not make the file smaller. It adds text. If the goal is fitting into an email attachment, the next step is Compress, with the estimated size shown before you pick a level. And when there is nothing to gain, the app tells you the PDF is already compact instead of handing back a worse file — the difference between compressing and pretending to.

After OCR: finding the file among hundreds#

Making a document searchable solves things inside the document. When the problem is “I know that invoice exists, I just don’t know which of the 4,000 PDFs it is in”, the tool is a different one: a content search that reads the files on your disk and answers as you type. That is what Elegant File Explorer does with deep search on, also 100% local — the search inside PDFs guide walks through it.

With both installed, they talk to each other: when the Finder matches a word inside a PDF, Elegant Paper opens at the right page and highlights the passage. It is a bonus, not a toll: each app solves its own problem on its own.