How to Extract Text from a Scanned PDF (Free & Paid Ways)

You open a PDF, drag across a paragraph, and nothing selects. That is the telltale sign of a scanned PDF: the “text” is actually a picture of text, so there is nothing for your cursor to grab. The fix is OCR — optical character recognition — which reads the picture and gives you real, copyable text.

Which method makes sense depends on how much you need:

• A few pages now and then → the free route with Capture2Text below takes a couple of minutes.
• An entire document (a book, a contract, a stack of invoices) → a dedicated tool that turns the whole file into a searchable PDF will save you hours.

Option 1: Free — Capture2Text, a Few Pages at a Time

Capture2Text is a free, open-source Windows OCR tool (GPLv3, portable, no install). It does not convert PDFs, but it is excellent at grabbing text from whatever is on your screen — including a PDF viewer. If you only need a page or two, this is the fastest free path.

1. Download Capture2Text (v4.6.3, portable) and run it. It sits quietly in your system tray.
2. Open the scanned PDF in your reader of choice and zoom in — larger text gives noticeably better OCR results. A zoom that fills the screen works well.
3. Press Win + Q (the default shortcut), drag a box around the paragraph you need, and release. The recognized text lands on your clipboard.
4. Paste it (Ctrl + V) into Word, Notepad, or wherever you need it.
5. Repeat for the next page or paragraph.

Honest limits: this is page-by-page work, so it does not scale to a 200-page document. It also will not create a searchable PDF — you get plain text in your clipboard, nothing more. Accuracy depends on scan quality: clean modern scans at high zoom are very good; faded, skewed, or handwritten pages will need manual cleanup.

Option 2: The Whole Document — Turn It Into a Searchable PDF

If you need the entire file converted, the right tool is one that runs OCR across every page and inserts an invisible text layer, producing a searchable PDF that looks identical but lets you select, copy, and search text. Capture2Text cannot do this — it is a screen-capture OCR tool, not a PDF converter.

Two established options for the job:

• Adobe Acrobat — the industry reference for “recognize text” in scanned PDFs. Reliable batch processing of multi-page scans.
• Wondershare PDFelement — a lower-cost alternative with a solid built-in OCR engine for turning scans into editable, searchable files.

The typical workflow in either: open the scanned PDF → choose the OCR / “Recognize Text” function → pick the document language → let it process → save. You end up with a PDF that looks the same but behaves like real text.

The Simple Rule of Thumb

Your situation	Best choice
You need a paragraph or a page from one PDF, occasionally	Capture2Text (free) — seconds per capture
You need the whole PDF searchable or editable	A PDF tool with OCR (Acrobat / PDFelement)
You receive scanned PDFs every week as part of your job	A PDF tool with OCR — the time savings pay for it quickly

Tips for Better OCR Results (Any Method)

• Zoom in / use a high-resolution scan. OCR engines read bigger, crisper glyphs far more accurately. If you can re-scan, 300 DPI beats 150 DPI clearly.
• Match the language. Capture2Text ships with English, French, German, Japanese, Korean, Russian, and Spanish built in; language packs for 90+ more languages are available from the download page. Wrong language selected = garbage output.
• Straighten skewed pages. A tilted scan is the most common cause of bad OCR. Most PDF tools can deskew before recognizing.
• Proofread the output. OCR is not 100%, even on clean scans. Numbers, proper names, and legal text deserve a second look — especially “l / I / 1” and “O / 0” confusions.
• Handwriting is a different game. Classic OCR (including Tesseract, the engine inside Capture2Text) is built for printed text. Handwritten notes will come out unreliable.