How to Convert a Scanned PDF to Word — OCR Step by Step
Direct answer: Go to convertfloor.com/tools/image-to-word. Upload your scanned PDF or image. Our OCR engine extracts the text and delivers an editable DOCX file. Free, no signup.
A scanned PDF is an image — a photograph of a document — rather than a real digital text file. Standard PDF converters cannot extract any text from it because there is no text to extract. To get editable Word output, you need OCR (Optical Character Recognition), which analyses the image and reconstructs the text.
This guide covers the fastest method, what to expect, and how to fix common OCR problems.
Step 1: Is Your PDF Scanned or Digital?
Before converting, confirm your PDF is actually scanned. There are two types:
Digital PDF
Scanned PDF
Created from Word, Excel, or design software
Created by scanning a paper document or photographing a page
Text is selectable — you can highlight it
Text is not selectable — it is part of the image
Standard PDF to Word converter works
Needs OCR first
Conversion is fast and accurate
Conversion accuracy depends on scan quality
Quick test: open your PDF in any viewer and try to highlight a word. If you can select text, it is digital. If you cannot, it is scanned.
Step 2: Convert Scanned PDF to Word Using OCR
Use convertfloor.com/tools/image-to-word — this converts directly from scanned image to editable DOCX in one step.
Go to Image to Word. Open convertfloor.com/tools/image-to-word.
Upload your scanned PDF. The tool accepts scanned PDFs, JPGs, PNGs, and photos of documents.
Wait for OCR processing. OCR takes longer than standard conversion — typically 15–60 seconds depending on file size and complexity.
Download your DOCX. The editable Word document downloads. Open it in Word, Google Docs, or LibreOffice.
Review and clean up. OCR is very good but not perfect. Review the output for errors before using the document.
What OCR Accuracy to Expect
OCR accuracy depends entirely on scan quality. Here is a realistic breakdown:
Scan Type
Expected Accuracy
Notes
Clear, straight, 300 DPI+ scan
98–99%
Near-perfect. Minor errors on unusual fonts.
Good quality but slightly skewed
90–95%
A few words wrong per page. Proofreading needed.
Phone photo (good lighting)
85–93%
Variable. Shadows and angles reduce accuracy.
Low-resolution scan (under 150 DPI)
60–80%
Many errors. Consider re-scanning at higher DPI.
Handwritten documents
20–60%
OCR is not designed for handwriting. Results vary greatly.
How to Improve OCR Accuracy Before Converting
1. Resolution — Aim for 300 DPI Minimum
DPI (dots per inch) is the single biggest factor in OCR accuracy. If your scan was done at 72 or 96 DPI (common on older scanners or phone photos), OCR will struggle. Re-scan at 300 DPI or higher if possible.
2. Straighten the Image
Skewed or rotated documents produce many OCR errors because the OCR engine reads line by line. Most scanners and phone scanner apps (like Adobe Scan, Microsoft Lens, or Apple's built-in scanner) straighten automatically. Use this before uploading.
3. Good Lighting (for Phone Photos)
If scanning with a phone camera, ensure even lighting with no shadows over the text. Photograph from directly above, not at an angle. Bright, indirect light works best.
4. Black and White Output
For text-only documents, scanning in black and white (not greyscale or colour) increases OCR accuracy and reduces file size. Colour scans contain more noise that confuses the OCR engine.
Fix Common OCR Errors After Converting
Extra Spaces and Line Breaks
OCR sometimes inserts extra spaces within words or line breaks in the middle of sentences. In Word: use Find & Replace (Ctrl+H) to find double spaces (' ') and replace with single space (' '). For paragraph breaks, use Find & Replace with the wildcard for paragraph marks.
Wrong Characters
Common OCR substitutions: 'l' (lowercase L) for 'I' (capital i) and '1' (one), 'O' (letter) for '0' (zero), 'rn' for 'm'. Always proofread numbers and names especially carefully.
Layout and Column Issues
Two-column documents sometimes have their columns merged into one stream of text. Re-arrange paragraphs manually or, for documents with complex layouts, OCR results may require significant manual reconstruction.
Alternative: Two-Step OCR Process
If you want to review the raw text before creating a Word document, use this two-step approach:
First, use convertfloor.com/tools/image-to-text to extract the text as plain text
Review and correct the text
Paste the corrected text into a new Word document and apply formatting
This takes more time but gives you full control over the output structure.