How to Extract Text From a PDF Image (Scanned Document)

When you receive a PDF that’s actually a scanned image, you can’t simply click and drag to select text. The document is a photograph of a page — the text is “baked in.” To get the text out, you need OCR.

Here’s every method, from simplest to most powerful.

---

Why Can’t You Copy Text From Some PDFs?

There are two types of PDFs:

1. Native/digital PDFs — created from Word, InDesign, etc. Text is stored as actual characters. You can search and copy freely.
2. Image-based/scanned PDFs — the page is stored as a raster image. No text data exists; just pixels.

If you try to Ctrl+A in a scanned PDF and no text gets selected, you have an image-based PDF.

---

Method 1: AltoUnlockPDF (Fast, Free, Online)

1. Visit our OCR tool
2. Upload your PDF
3. Select output format: Searchable PDF (keeps original appearance + adds text layer) or Plain Text (.txt)
4. Choose language
5. Download output

Takes about 5–30 seconds per page. No signup required.

---

Method 2: Google Drive (Free, Highly Accurate)

1. Upload PDF to Google Drive
2. Right-click → Open with → Google Docs
3. Wait 30–60 seconds for OCR to complete
4. The document opens with extracted text above/below each page image
5. Select all → Copy → paste wherever needed

Works great for 1–20 page documents. Free and unlimited.

---

Method 3: Python — Programmatic Extraction

For developers or bulk processing:

import pdf2image
import pytesseract
from PIL import Image
import io

def extract_text_from_scanned_pdf(pdf_path):
    # Convert PDF pages to images
    images = pdf2image.convert_from_path(pdf_path, dpi=300)
    
    text_pages = []
    for i, image in enumerate(images):
        # Run OCR on each page
        text = pytesseract.image_to_string(image, lang='eng')
        text_pages.append(f"--- Page {i+1} ---\n{text}")
        print(f"Processed page {i+1}/{len(images)}")
    
    return '\n\n'.join(text_pages)

# Usage
text = extract_text_from_scanned_pdf('contract.pdf')
with open('contract_text.txt', 'w') as f:
    f.write(text)

Dependencies:

pip install pdf2image pytesseract Pillow
# Also install: poppler (for pdf2image) and tesseract (for pytesseract)

---

Method 4: Adobe Acrobat Reader (Recognize Text)

Even the free Adobe Acrobat Reader can recognize text in scanned PDFs:

1. Open the scanned PDF in Adobe Acrobat Reader
2. Look for the notification bar: “This document contains only images”
3. Click “Recognize Text” (appears in the right panel or notification)
4. Wait for processing
5. Now Ctrl+F search and text selection work

Limitations: free version can recognize but may not let you export the text layer.

---

Method 5: macOS Preview (Built-In on Mac)

macOS Preview has improved significantly and now includes basic OCR:

1. Open the scanned PDF in Preview
2. Select the text tool (T)
3. Try to click and drag on text areas
4. If OCR is needed, use Edit → Redact or import to Notes for Apple Intelligence OCR

Apple’s Live Text feature in macOS Monterey+ recognizes text in images automatically when you use the selection tool.

---

Extracting Text While Preserving Formatting

Sometimes you need the text with its original structure (columns, tables, headings). Tools for this:

• ABBYY FineReader (paid) — best structure preservation
• Adobe Acrobat Pro (paid) — good table extraction
• Camelot (Python, free) — specifically for tables in PDFs:

import camelot
tables = camelot.read_pdf('annual_report.pdf', pages='1-5')
tables[0].df  # Returns a pandas DataFrame
tables.export('tables.csv', f='csv')

---

Which Method Should You Use?

Scenario	Best Method
Quick, one-time extraction	AltoUnlockPDF or Google Drive
Bulk processing (100+ PDFs)	Python + pytesseract
Mac user	macOS Live Text + Preview
Need tables preserved	ABBYY FineReader or Camelot
Developer building a product	pytesseract API or cloud OCR API

The best free combination for most people: Google Drive for single documents, Python script for batch processing.