OCR stands for Optical Character Recognition. In plain terms, it takes an image that contains text — a scan, a photo, a screenshot — and turns the shapes of the letters back into real text a computer can search, copy and edit.
The problem OCR solves
When you scan a page, the computer sees a picture. To it, the text is just patterns of pixels — no different from a photo of a landscape. You cannot search it, copy it, or read it with a screen reader.
OCR looks at those pixel patterns and recognises the letters, then reconstructs them as actual characters.
- Scanned contracts become searchable
- Photos of receipts become expense data
- Old paper records become editable text
How it works, simply
Modern OCR runs an image through a model trained on millions of examples of text. It detects lines, breaks them into characters, and predicts which letter each shape most likely is. The best systems also understand layout, so headings, columns and tables survive.
OCR is translation — from a picture of a word, to the word itself.
Why it matters more than ever
OCR sits underneath a surprising amount of software: searchable archives, automatic data entry, document AI, and accessibility tools that read documents aloud. It is the bridge between the paper world and the digital one.
Try it with SELKATDocs
SELKATDocs runs OCR on scanned PDFs and images right in your browser, so you get editable text without uploading anything to a stranger's server.
Once text is real text again, everything else — search, copy, edit, sign — becomes possible.