A scanned PDF is essentially a stack of images inside a PDF wrapper. You can't select text, you can't search it, and you can't feed it into other tools that expect real text. OCR (Optical Character Recognition) fixes that by reading the images and producing actual characters.
When you need OCR
If Ctrl-F / Cmd-F finds nothing in a PDF, or if you try to select text and end up drawing a rectangle over an image, the PDF has no text layer. That's when OCR is required.
Steps
1. Open /ocr-pdf.
2. Upload the scanned PDF.
3. Pick the language — English, Hindi, Spanish and French are supported.
4. Press OCR PDF. Processing takes a few seconds per page.
5. Copy the extracted text or download it.
Tips for higher OCR accuracy
Scan at 300 DPI or higher. Below 200 DPI, OCR quality drops sharply.
Black-and-white or high-contrast scans work best. Colour photos with shadows confuse the recognizer.
One language at a time. Mixed-language documents produce worse results than pure single-language pages.
Straighten the scan. Even a 5° tilt hurts accuracy. Use Rotate PDF to fix obvious tilts.
What OCR cannot do
OCR is designed for printed text. Handwriting is unreliable — legible printed handwriting works occasionally, but casual handwriting is a lottery. Mathematical formulas often come out garbled. Complex multi-column layouts are usually flattened into a single reading order.
After OCR
You can pipe the OCR-ed PDF into other tools: PDF to Word for editing, Translate PDF for a different language, or AI Summarizer for the key points.
OCR is one of those tools you don't need often, but when you do, nothing else works. A clean scan + the right language pick + a couple of minutes of processing is usually enough.