How to Make a Scanned PDF Searchable (OCR Explained)

You scan a contract, and a month later you need the one clause about a notice period. You press Ctrl+F, type the word, and get nothing - even though you can plainly see it on screen. The document looks like text but the file contains none. This guide explains why that happens, how OCR fixes it with an invisible text layer, and how to turn a scan into a PDF you can search, select and copy from.

Why a scanned PDF contains no text

A PDF is a container, and it can hold two very different kinds of page. One exported from Word or a browser stores real text objects: character codes, a font, and coordinates for each glyph. Search works because the file literally contains the letters.

A scanned PDF holds none of that. Your scanner or phone camera produced a photograph of the paper, and the PDF wraps that photograph in a page. To your eyes the result is words; to software it is a grid of light and dark pixels. Nothing is broken and nothing is hidden - the text was never encoded in the first place. That is why search finds nothing, why you cannot select a sentence, and why PDF to Word comes back empty on a scan.

How to tell whether your PDF needs OCR

The test takes five seconds. Open the PDF and try to drag-select a line of text:

• If a normal blue text highlight follows the words, the PDF already has a text layer and does not need OCR.
• If you get a rectangular marquee - or nothing at all - the page is an image and needs OCR.
• Mixed documents are common: a typed report with a signed, scanned page on the end. Only the scanned pages lack text.

File size is a second clue. Ten pages of real text is often under 200 KB; ten scanned pages is usually several megabytes, because each one is a full-page image.

What OCR adds: the invisible text layer

OCR - optical character recognition - looks at the pixels, works out which shapes are which characters, and reports both the words and the coordinates of the box each one occupies. That second part is what makes a searchable PDF possible.

Rather than replacing your scan with retyped text, the OCR PDF tool keeps the original page image and draws the recognised words on top in an invisible font, each positioned over the pixels it came from. This text-layer sandwich has a useful property: the page still looks identical, so its appearance cannot go wrong, but Ctrl+F now finds words and copy-paste yields real characters. If OCR misreads a word, only the hidden copy is wrong - what you see is still the original.

How to make a scanned PDF searchable

1. Open the OCR PDF tool and drop your scanned PDF onto the page.
2. Pick the language in the document - English, Hindi, or English + Hindi for a mixed one.
3. Click to start. The first run downloads the engine and language data, a few megabytes, then caches it.
4. Wait while each page is rendered and read in turn. Progress is shown per page, because OCR is CPU-intensive.
5. Your browser saves the result as your-file-searchable.pdf. Open it and press Ctrl+F to confirm.

A worked example

Say you have a 12-page scanned tenancy agreement and need every mention of "deposit". Each page is rendered at roughly 144 DPI, so an A4 page - 8.27 by 11.69 inches - becomes an image about 1,191 by 1,684 pixels: enough detail to separate an e from a c at normal type sizes, without images so large the output PDF becomes unwieldy.

The pages are read one after another, then reassembled into a new PDF with the text layers attached. Search the result for "deposit" and the reader jumps straight to the clauses that mention it, highlighting each in place on the scan. You can copy the amount into an email instead of retyping it - and, just as usefully, PDF to Word or PDF to Excel will now produce content, because those tools extract a text layer and there is finally one to extract.

What affects accuracy

OCR quality is decided mostly before you upload anything - by how the scan was made:

• Resolution. Scan at 300 DPI where you can. Small print at 150 DPI leaves the engine guessing between similar shapes.
• Straightness. A page skewed by even a few degrees hurts noticeably, because line detection assumes roughly horizontal rows. Straighten the paper, not the file.
• Contrast. Crisp black on white is ideal. Phone photos with a shadow across the page, or a grey cast, cost accuracy - most scanner apps have a document mode that fixes both.
• Typeface. Clean printed type reads best. Handwriting, script fonts and dot-matrix print are far less reliable, and handwriting should not be relied on at all.
• Language. Selecting the right language loads its recognition model and character set. Choosing English for a Hindi document returns nonsense.

OCR is very good but never perfect. For anything consequential - an amount, a date, an account number - read it off the page image before relying on it.

Things worth knowing before you run it

• Only run it on scans. Each page is re-rendered as an image, so a PDF that already has real text would come back rasterised - visually similar but no longer true text. Do the select test first.
• Expect it to take a while. Every page is rendered and then read on your own device, so a hundred-page scan is a coffee-break job rather than an instant one.
• Output size can grow, because the copy stores a freshly rendered image per page. If it needs to go by email, follow up with Compress PDF.
• Password-protected files must be opened first - remove restrictions with Unlock PDF, then try again.

Images versus PDFs

If what you have is a photo or screenshot - a receipt, a whiteboard, a page of a book - you probably want the text itself rather than a searchable document, and extract text from an image gives you plain text to paste anywhere. Choose OCR PDF when you want the document to stay a document: same pages, same signatures and stamps, but findable.

Why it matters that this runs in your browser

Scanned documents are, by their nature, the sensitive ones: contracts, ID pages, medical letters, bank statements. Most online OCR services upload your file to a server, so a copy sits on someone else's machine under their retention policy. The OCR PDF tool instead does the whole job inside your browser tab - nothing is transmitted or stored, and it keeps working offline once the engine is cached, which is a fair sign no server is involved.

Make your scanned PDF searchable

If Ctrl+F comes up empty on a document you can clearly read, the file simply has no text in it yet. Open the OCR PDF tool, drop the scan in, choose the language, and download a copy that looks identical but can be searched, selected and copied - free, and without the document leaving your device.