How to Clean Up Text Copied from a PDF

This regex matches a newline that isn't preceded by another newline and isn't followed by another newline - in other words, a single line break. It replaces it with a space, joining the broken lines back together. Double newlines (paragraph separators) are left alone.

In Google Docs:

Google Docs doesn't support regex in find-and-replace, so the process is more manual. Paste your text, then use Edit > Find and Replace. Search for the paragraph mark (you can't type it directly, but you can copy a line break from the document and paste it into the search field). Replace with a space. Then manually re-add paragraph breaks where they belong.

Fixing extra spaces

Multiple spaces between words are easy to fix with find-and-replace. Search for two spaces and replace with one. Repeat until no more double spaces remain. Or use regex:

Find:    [ ]{2,}
Replace: [space]

This collapses any run of 2+ spaces into a single space in one pass.

Fixing hyphenated words

For hyphens at line breaks, search for a hyphen followed by a newline and replace with nothing - this rejoins the word:

Find:    -\n
Replace: [empty]

Be careful with this one. It works perfectly for words like "environ-\nment" becoming "environment," but it'll also join intentionally hyphenated words that happen to fall at a line break. Read through the result to check for words that got incorrectly merged.

Command-Line Tools for Better PDF Text Extraction

If you deal with PDFs regularly, there are tools that extract text more cleanly than copy-paste does. They analyze the PDF's internal structure rather than scraping the rendered view.

pdftotext (Linux/Mac)

Part of the poppler-utils package. It reads the PDF's internal text streams directly, which usually produces cleaner output than clipboard copying.

# Basic extraction
pdftotext document.pdf output.txt

# Preserve layout (keeps columns, tables)
pdftotext -layout document.pdf output.txt

# Extract just pages 5-10
pdftotext -f 5 -l 10 document.pdf output.txt

# Output to stdout (pipe to other tools)
pdftotext document.pdf -

Install it on Ubuntu/Debian with sudo apt install poppler-utils. On Mac, use brew install poppler.

Python: pdfplumber

For more control, the Python pdfplumber library gives you access to character-level positioning data, making it possible to reconstruct paragraphs intelligently.

import pdfplumber
import re

with pdfplumber.open("document.pdf") as pdf:
    for page in pdf.pages:
        text = page.extract_text()
        # Join broken lines
        text = re.sub(r'(?<!\n)\n(?!\n)', ' ', text)
        # Fix double spaces
        text = re.sub(r' {2,}', ' ', text)
        print(text)

Install with pip install pdfplumber. It handles tables, multi-column layouts, and cropped regions better than most alternatives.

Scanned PDFs: A Different Problem Entirely

Everything above assumes your PDF contains actual text data. But some PDFs are just images of text - scanned documents, photos of pages, or PDFs exported from certain older systems. When you try to copy text from these, you either get nothing at all, or you get a jumble of random characters from a failed OCR attempt.

You can tell the difference by trying to select individual words. If you can highlight word by word, the PDF has real text. If the selection covers a rectangular block regardless of word boundaries, it's a scanned image.

For scanned PDFs, you need OCR (Optical Character Recognition) to convert the image to text first. Free options include:

• Google Drive - Upload the PDF to Google Drive, then open it with Google Docs. Google runs OCR automatically and gives you an editable document.
• Tesseract - Open-source OCR engine. Run tesseract input.png output on the command line. Best results with high-resolution, clean scans.
• Adobe Acrobat - The "Recognize Text" feature runs OCR and creates a searchable PDF layer. Paid software, but very accurate.

After OCR, you'll still need to clean up the extracted text - OCR output tends to have its own quirks, like confusing "l" with "1", "O" with "0", or dropping punctuation. Paste the OCR output into Article Formatter to fix the line break and spacing issues, then proofread for character recognition errors.

Tips for Cleaner PDF Copies

A few habits can reduce the cleanup work before you even paste:

Multi-Column PDFs

Two-column layouts (common in academic papers and newsletters) are especially tricky. When you copy text from a two-column PDF, the copied text often interleaves lines from both columns:

The experiment showed that    Results were consistent
participants preferred the     across all three trial
simpler interface design,      groups, with a margin
achieving task completion      of error below 2%.

This happens because the PDF reader scans left to right across the full page width, grabbing text from both columns on the same line. The fix depends on the tool you're using:

• Manual selection: Carefully select just one column at a time. In Adobe Reader, hold Alt while dragging to select a rectangular region.
• pdftotext with -layout: The -layout flag preserves the spatial arrangement, making it easier to separate columns afterward.
• pdfplumber: You can define crop regions to extract each column separately: page.within_bbox((0, 0, page.width/2, page.height)) for the left column.

When Nothing Works

Some PDFs are genuinely impossible to copy text from cleanly. Heavily designed documents with text baked into graphics, PDFs with DRM restrictions that prevent copying, and very old PDFs with non-standard encodings can all resist extraction.

In these cases, your options are:

• Contact the source. Ask for the content in another format. Most authors or organizations will provide an alternative if asked.
• Screenshot and OCR. Take a screenshot of the PDF page, then run it through Google Drive or Tesseract OCR. This bypasses the PDF's internal encoding entirely.
• Manual retyping. For short passages, it's sometimes faster to just type it out. Use the Word Counter to track your progress as you go.

Frequently Asked Questions

Related Tools