Why you can’t select text in a scanned PDF, what OCR does about it, and how to get clean results.
You scan a contract, a form or an old book page. The PDF looks perfect on screen. But when you try to search for a word, nothing is found. When you try to copy a sentence, you can only drag a box over a picture.
That’s because a scanned PDF isn’t really text. It’s a photograph of a page. The fix is a technology called OCR, and once you understand it, you can turn almost any scan into a document you can search, copy and edit.
Why a scanned PDF has no text
A normal PDF, such as one exported from Word, stores the actual letters of every word. A scanned PDF stores a picture of each page, made of tiny coloured dots. Your eyes see words, but the computer sees only an image.
That’s why you can’t search it, copy from it, convert it cleanly to Word, or have a screen reader read it aloud.
An easy test: open the PDF and try to select one sentence with your mouse. If it highlights, the text is real. If you can only select the whole page or draw a box, it’s a scan.
What OCR does
OCR stands for optical character recognition. It looks at the picture of each page, finds the shapes of letters, and works out which characters they are. The result is real text, which is usually placed invisibly behind the scanned image, so the page looks the same but you can now search and copy it.
That gives you:
Searchable PDFs. Find any word with Ctrl+F.
Copyable text. Select a paragraph and paste it into another document.
Better conversions. A scan with OCR converts to Word or Excel far better than one without.
Accessibility. Screen readers can read the document aloud.
When you need it
Scanned contracts, letters and forms
Old documents and archives you want to search
Receipts and invoices for expense reports
Pages from a book or a report that you want to quote
A scanned PDF that you plan to convert to Word
Your options for OCR
Use a desktop program. Many PDF programs include OCR. The best ones are often paid, and the free ones can be limited.
Use a phone scanning app. Some apps run OCR as you scan. It’s convenient for new documents, but not for files you already have.
Use an online OCR tool. You upload the scanned PDF, run OCR and download the result. Nothing is installed, and it works on any device. The trade-off is that your document is uploaded, so choose a tool that explains how it handles files.
How to run OCR on a PDF online, step by step
The steps are similar in most online tools. Here they are using the one I built, MyPDFLabs.
Upload your scanned PDF. Drag it into the box or click to browse. The tool handles one file at a time, and it shows the file name and size once it’s added.
Choose the document language. English is selected by default. Pick the main language of your document, because the tool says this improves accuracy.
Click Run OCR. The tool notes that it works best with clean, high-resolution scans, and that processing may take 1 to 3 minutes. A long or large document can take longer, since every page has to be analysed.
Download the result and save it under a new name, such as contract-searchable.pdf.
Test it. Open the file, press Ctrl+F, and search for a word you can see on the page. Then try copying a sentence.
Proofread the important parts. OCR is good, but it isn’t perfect, so check names, numbers and dates.
How to get better OCR results
The quality of the scan matters more than the tool. A few habits make a big difference:
Scan at a good resolution. Around 300 dpi is a common recommendation for text documents.
Keep the page flat and straight. A tilted or curved page, like a book spine, is harder to read.
Use good light and contrast. Dark text on a clean, light background works best.
Avoid shadows and blur. For a phone photo, hold the phone steady and photograph the page from directly above.
Fix orientation first. A sideways page can confuse OCR, so rotate it before you run it.
Choose the right language. OCR is more accurate when it knows the language of the text, so set it before you start.
Keep the original scan. Save the OCR result as a new file.
What OCR struggles with
Handwriting. Neat printed handwriting sometimes works, but messy writing often doesn’t.
Very small or faded text. Old photocopies and light print can lose letters.
Complex layouts. Columns, tables and text over images can end up in the wrong order.
Look-alike characters. The digit 0 and the letter O, or 1, l and I, are easy to mix up.
Unusual fonts and decorative text. The more stylised the letters, the more mistakes.
That’s why proofreading matters, especially for anything with numbers or legal wording.
Common problems and how to fix them
The text I copied is full of strange characters. The scan was probably blurry or low resolution. Rescan at a higher quality and run OCR again.
Some words are wrong. Check that you chose the right language. Then correct the words by hand, or rescan the pages that went badly.
The text comes out in the wrong order. The page has columns or a complex layout. Check the order, or work on one column at a time.
The file got much bigger. OCR adds a text layer, and the scan itself can be large. Compress the PDF afterwards.
Nothing was recognised. The page might be sideways, upside down, or too faint. Rotate it, or rescan with better contrast.
It’s taking a long time. Large files take longer. Give it a few minutes before you assume something went wrong.
It’s a handwritten document. Expect poor results, and plan to type the important parts yourself.
Habits that save you trouble
Keep the original scan. If something goes wrong, you can run OCR again.
Proofread numbers and names. A single wrong digit in a contract or an invoice matters.
Test one page before you run a big document. Make sure the quality is good enough before you wait for 200 pages.
Be careful with sensitive files. Before you upload a contract, ID or financial paper to any online tool, read what it says about storing and deleting files. If you’re not comfortable, run OCR on your own computer.
A free OCR tool I built
Full disclosure: I built MyPDFLabs, a free set of online PDF tools. Its OCR tool turns scanned PDFs into searchable documents, with a choice of document language, no sign-up and no watermark on the core tools. It handles one file at a time, supports files up to 500MB, and deletes uploaded files automatically after processing. Results depend on the quality of your scan. It’s my own project, so treat it as one option among many. The advice in this article works with any OCR tool.
The short version
A scanned PDF is a picture, so it has no searchable or copyable text.
OCR reads the picture and turns it into real text.
Clean, straight, well-lit scans at around 300 dpi give the best results, and choosing the right language helps.
Always proofread names, numbers and dates, and keep the original scan.
Think before uploading sensitive documents anywhere.
Thanks for reading. If there’s a PDF task that always slows you down, tell me in the comments and I’ll cover it next.
