Can I make a scanned PDF searchable without changing how it looks?

Yes, and that is the normal way OCR is applied. The recognised text is placed in an invisible layer positioned behind the original image, so the page you see is untouched while search, selection, and copying all start working.

My PDF is partly selectable and partly not. What is going on?

It is a mixed file — some pages were generated as text and others were scanned and inserted, which is very common in signed contracts where only the signature page went through a scanner. Running OCR over the whole document is harmless: pages that already have text are left alone.

Does OCR change my original text?

No. OCR adds a layer rather than replacing anything. If the recognition is poor the underlying page is still exactly the scan you started with.

how to tell if a pdf is scanned

Scanned PDF or Text PDF? How to Tell

Try to select a word. If the cursor picks out the word, you have text. If it draws a box over the whole page, you have a photograph — and that single fact explains most of what a PDF will and will not let you do.

The two kinds

A text PDF stores characters. It knows the page contains the word "invoice", it knows which font to draw it in, and it knows exactly where on the page it goes. Anything produced by Word, a browser, an accounting package, or an export from almost any program is this kind.

A scanned PDF stores pixels. A scanner, a phone camera, or a fax has photographed a sheet of paper and wrapped the resulting image in a PDF container. The file has no idea there are words on it. To the software it is one large picture that happens to be page-shaped.

Both open in the same viewer and look much the same on screen. Everything else about them differs.

How to check

Select a word with your cursor. Text highlights word by word and follows the line. A scan gives you a rectangle over the whole image, or nothing.

Search for a word you can see. Ctrl+F, type it, press enter. No result on a word plainly visible on the page is conclusive: there is no text layer.

Look at the file size. Twenty pages in 200 KB is text. Twenty pages in 25 MB is scans.

Zoom in hard. Text stays crisp at 400% because it is redrawn at every zoom level. A scan turns into visible pixels or JPEG mush.

Why it matters

It decides what is possible. You cannot search, copy, or convert a scan to Word in any meaningful sense, because there is nothing in the file to search, copy, or convert. Tools that appear to do it are running OCR first, whether or not they say so.

It also decides file size, as above, and it decides accessibility: a screen reader can read a text PDF aloud and can make nothing at all of a scan. If a document has to be accessible, this is the first thing to check, not the last.

Search and copy: Text only. A scan needs OCR before either works.
Convert to Word or Excel: Text converts directly. A scan converts only as well as OCR reads it.
File size: Text is small and stays small. Scans are large and only compression helps.
Screen readers: Text can be read aloud. A scan is silent.

Turning a scan into text

OCR — optical character recognition — looks at the picture and works out which letters it is seeing. Run over a scanned PDF, it adds an invisible text layer behind the image, so the page looks exactly as it did but is now searchable, selectable, and convertible.

It is a reading process, not a conversion, so it is never perfect. Clean, straight, 300 DPI printed text is read very accurately. Faint photocopies, tight columns, unusual fonts, and handwriting are read badly or not at all. Check the result before you rely on it, particularly for numbers, where a misread digit looks entirely plausible.