Online Tool Store
· 4 min read
How to Extract Text From a Scanned PDF
Heshan Fernando
Co-founder & COO
You try to select text in a PDF — to copy a paragraph, or search for a specific term — and nothing happens. The document is a scanned contract, an old book, or a faxed form: as far as your computer is concerned, it’s just a picture of text, not actual text. This trips people up constantly, because a scanned PDF looks exactly like a normal, text-based one until you try to interact with it.
The difference matters more than it seems. A text-based PDF is searchable, copyable, and accessible to screen readers; an image-only scanned PDF is none of those things until OCR (optical character recognition) processes it and recovers the actual text underneath the picture.
What OCR on a scanned PDF actually involves
A scanned PDF is, internally, just a sequence of full-page images — one per page — with no underlying text layer at all. OCR analyzes each page image, identifies regions that look like text, and matches those shapes against known letterforms to reconstruct the actual words. The output is genuine, selectable text that corresponds to what’s visually on the page.
This works page by page because each page is effectively its own image — a multi-page scanned document means running the same recognition process across every page and assembling the results in order. Accuracy depends heavily on scan quality: a clean, high-contrast, well-aligned scan gives much better results than a skewed, low-resolution, or poorly lit one.
Why people get stuck here
- A scanned PDF looks normal until you try to interact with it. There’s no visual cue that a PDF lacks a text layer — you only discover it when copy or search fails.
- Old or faxed documents are especially common candidates. Contracts, historical records, and faxed forms are frequently scanned-only, with no original digital text version available.
- Manual retyping doesn’t scale. Retyping a multi-page scanned document by hand is realistic for a page, not for a whole document.
- Not everyone realizes OCR is the actual solution. People sometimes assume a “broken” PDF just can’t be searched or copied, without knowing OCR can recover that functionality.
What a good PDF OCR tool looks like
Processes every page, not just one
A useful tool should handle a full multi-page scanned document, running recognition across every page and assembling the results.
Handles imperfect scan quality reasonably well
Real-world scans aren’t always perfectly aligned or high-resolution — good OCR tolerates a reasonable amount of skew, noise, and imperfect lighting.
Keeps the document private
Scanned documents are often exactly the kind of content — contracts, personal records — people are least comfortable uploading to an unfamiliar server for processing.
Common mistakes to avoid
- Assuming a scanned PDF simply “can’t” be searched or copied, when OCR can recover that functionality from the page images.
- Running OCR on a low-quality, heavily skewed scan and expecting perfect results — accuracy depends significantly on scan quality.
- Skipping a proofread of OCR output for an important document, especially around numbers, which are more prone to misreads than plain text.
- Uploading a sensitive scanned document (a signed contract, a medical record) to an unfamiliar OCR service without checking how it handles your data.
How to do it with the PDF OCR Tool
Online Tool Store’s PDF OCR Tool recognizes text from a scanned PDF entirely in your browser.
- Upload your scanned or image-only PDF.
- Let the OCR engine process each page.
- Review the recognized text for accuracy, especially numbers and unusual formatting.
- Copy or export the resulting text for use in a document, search, or archive.
Because nothing is uploaded to a server, it’s a reasonable choice even for scanned contracts and other sensitive documents you’d rather keep off a third-party service.
Frequently asked questions
How can I tell if a PDF is scanned versus text-based?
Try selecting text on the page — if you can highlight and copy words normally, it’s text-based. If clicking and dragging just selects the whole page as an image with no individual words highlighting, it’s a scanned, image-only PDF that needs OCR to become searchable.
How accurate is OCR on an old or faded scanned document?
Accuracy depends heavily on scan quality — clean, high-contrast, well-aligned scans of even old documents can OCR quite accurately, while faded, skewed, or low-resolution scans produce more errors. Reviewing the output against the original is worthwhile for anything important.
Does OCR preserve the original document’s formatting?
Basic OCR recovers the text content but doesn’t always perfectly preserve complex layout (multi-column text, tables) — the recognized text may need some manual cleanup for documents with unusual formatting, even though the core word recognition is accurate.
Final thought
A scanned PDF isn’t actually stuck — the text is there, visually, on every page; OCR just does the work of turning that picture of text back into text you can search, copy, and quote.