What is OCR, and when do you actually need it?

You search a PDF for a name you know is in there, and nothing comes up. The text is right on the screen. You can see it. But the search box acts like the document is empty. That gap is the whole reason OCR exists.

Two kinds of PDF that look identical

A PDF can hold text in two completely different ways, and on screen they can look the same.

The first kind is a real text PDF. It was exported from Word, a website, an accounting app, something digital. The letters are stored as actual characters. The PDF knows the word “invoice” is sitting in the top corner. You can select it, copy it, search it.

The second kind is a scanned PDF. Someone put a paper document on a scanner, or took a photo of it with their phone, and saved that picture inside a PDF. To you it reads fine. To the computer it is just a photo. There are no letters in there, only colored pixels arranged in the shape of letters. Search finds nothing because there is no text to find.

The quick test: try to select a single word with your cursor. If you can highlight it, the text is real. If your cursor draws a box over the whole page like it is an image, you have a scan.

What OCR actually does

OCR stands for Optical Character Recognition. It looks at the picture, finds the shapes that are letters, and writes out the actual text those shapes represent. Then it tucks that text into the PDF, behind the image, lined up with what you see.

The page still looks exactly the same. The scan, the coffee stain, the slightly crooked angle, all of it stays. But now there is a layer of real text underneath. So search works. Copy works. The document went from being a photo of words to being words.

When you actually need it

Most of the time you do not. A PDF you exported yourself already has real text. Running OCR on it would be pointless. You need OCR when the text is locked inside an image, and that happens more than you would think.

Searching old scanned files. You have a folder of receipts or invoices, all scanned years ago, and you need the one from a specific vendor. Without OCR you open them one by one. With OCR you search the folder and find it in a second.

Copying text from a photographed document. Someone sends you a contract as a phone photo. You need to quote a clause in an email. Retyping it by hand is the slow path. OCR lets you select that paragraph and paste it.

Making a document accessible. A scanned PDF is a wall to anyone using a screen reader. The software reads the text layer aloud, and a pure image has no text layer, so it reads nothing. OCR gives it something to read. This is also why scanned documents fail accessibility checks for things like web publishing or public records.

Letting AI read the document. If you want to summarize a scanned report or ask questions about it, the tool needs text to work with. A raw scan gives it pixels and a guess. A document with a real text layer gives it the actual words, and the answers get a lot better.

How files.co handles it

The OCR on files.co runs with Tesseract, a well-known open source engine, and it runs inside your browser. Your document is never uploaded. It does not touch a server. The recognition happens on your own machine, on the page you already have open, and the file with its new text layer is built right there for you to download.

That matters because the documents people most often need to OCR are the private ones. Scanned contracts, medical letters, bank statements, ID copies. Sending those to a stranger’s server to be read is a strange tradeoff for searchable text. Here you do not make it. The page does the work and the file stays with you.

You can try it with our free OCR tool.

The short version

If you can select the words in a PDF, the text is already there and you are done. If your cursor only draws a box over an image, the words are trapped in a picture, and OCR is the tool that sets them free. Searchable, copyable, readable by a screen reader, readable by AI. Same page you started with, now with the text it was always missing.