You copy three paragraphs out of a PDF, paste them into an email or a document, and get a mess. Lines break in the middle of sentences. Words are split with stray hyphens. Page numbers and running headers appear halfway through a paragraph. Sometimes “fi” and “fl” vanish from words entirely, or quotation marks turn into odd symbols.
None of this is your fault. A PDF doesn’t store paragraphs the way a word processor does. It stores instructions for where each character or word should be drawn on a page. When you copy, your computer has to guess where sentences begin and end, and that guess often goes wrong, especially with columns, justified text and scanned pages.
The good news is that almost every problem follows a pattern, and patterns can be fixed. Here are eight ways to get clean, usable text out of a PDF, roughly in the order you should try them.
1. Check whether the PDF actually contains text
Before you fix anything, find out what you’re working with. Open the PDF and try to highlight a single word. If the cursor selects text, the file has a real text layer. If it draws a box over part of the page instead, or selects the whole page as one block, you’re looking at an image of text, usually a scan or a photo.
Copying from an image gives you nothing, or gibberish. In that case you need optical character recognition (OCR) first. Many PDF readers and scanner apps include OCR, and it’s worth running before you try any of the fixes below. Expect to proofread the result: OCR confuses similar shapes, such as “rn” and “m”, “0” and “O”, or “1” and “l”.
2. Export the text instead of copying it
Copy and paste is the least reliable way to get text out of a PDF. A dedicated export reads the whole text layer in order and usually keeps paragraphs together far better than your clipboard does.
You have two good options. If you only need words, convert the file to a plain .txt file. If you need to keep headings, lists and basic formatting, convert it to an editable Word document instead and copy from there. Free tools such as PDFVerge do both in the browser, with PDF to TXT and PDF to Word converters, so you don’t have to install anything.
Exported text still needs a quick tidy, but you’ll start with far fewer broken lines.
3. Trim and straighten the PDF before you extract
Extraction is cleaner when the file is cleaner. Three small jobs make a big difference:
Remove pages you don’t need. If you only want chapter four, split it out or delete the other pages first. You’ll have less to clean up, and fewer stray headers and footers.
Rotate sideways pages. Landscape tables and pages that were scanned at an angle often come out in a scrambled order. Rotate them upright before you export.
Watch for multi-column layouts. Newsletters and academic papers set in two columns are the worst offenders, because some tools read straight across both columns line by line. If that happens, copy one column at a time, or split the page into the parts you need.
4. Rejoin lines that break mid-sentence
The most common problem is a hard line break at the end of every line, exactly where the line ended on the PDF page. Paragraphs become ragged stacks of half-sentences.
The fix is to remove single line breaks while keeping the blank lines between paragraphs. Most code editors, including VS Code, Notepad++ and Sublime Text, support regular expressions in Find and Replace. Turn regex on, then:
Find: (?<!\n)\n(?!\n)
Replace with: a single space
That pattern matches a line break that isn’t next to another line break, which means it joins lines inside a paragraph but leaves paragraph breaks alone. (In Notepad++ on Windows, convert the line endings to Unix format first, via Edit > EOL Conversion, so the pattern sees plain \n breaks.) If you’d rather not use regex, paste the text into a plain text converter or a “remove line breaks” tool that has an option to keep paragraph spacing.
5. Repair words split by hyphens
Justified PDFs often hyphenate long words at the end of a line: “docu-” on one line, “ment” on the next. After you join the lines, you’re left with “docu- ment” or “docu-ment”.
Search for a hyphen followed by a line break (-\n with regex on) and replace it with nothing, before you run step 4. Then skim the result for genuine compound words, such as “well-known” or “follow-up”, that happened to fall at a line end. A spell checker is the fastest way to catch both kinds of mistake.
6. Strip invisible characters and fix spacing
Some of the strangest problems come from characters you can’t see:
Non-breaking spaces look like normal spaces but stop text from wrapping and can break search, formulas and code.
Soft hyphens are invisible until the text reflows, then appear in the middle of words.
Zero-width spaces can make two identical-looking words fail to match.
Tabs and double spaces come from justified text and table layouts.
A plain text converter or a “clean text” tool can remove these in one pass. In a regex editor, replace \x{A0} (a non-breaking space) with a normal space, delete \x{AD} and \x{200B}, and replace runs of spaces ([ ]{2,}) with a single space. Some editors write these as \u00A0, \u00AD and \u200B instead.
7. Fix ligatures, quotes and symbols
Typeset PDFs often use ligatures: single characters that combine letters such as “fi”, “fl” and “ff”. When copied, a ligature may paste as one unusual character, or drop out completely, leaving “eective” instead of “effective” and “rst” instead of “first”.
Search for the ligature characters (fi, fl, ff, ffi, ffl) and replace them with the plain letters. While you’re there, decide what you want to do with curly quotes, apostrophes and long dashes. They’re fine in a document, but if the text is going into code, a spreadsheet, a database or a web form, convert them to straight quotes and plain hyphens. Bullet symbols are another frequent casualty: “•” can arrive as a box, a question mark or a stray letter.
8. Remove headers, footers and page numbers
Running headers (“Annual Report 2025”), footers (“Confidential”) and page numbers get extracted along with the body text. They land in the middle of paragraphs at every page break.
Because they repeat, they’re easy to find. Search for the header text and remove every instance. Page numbers usually sit on a line of their own, so a regex such as ^\d+$ (with multiline matching) finds lines containing only a number. Check before you delete all of them, in case a table or list contains numbers on their own lines.
A quick order of operations
When you have a long document to clean, the order matters, because some fixes depend on others:
Confirm the PDF has a text layer, and run OCR if it doesn’t.
Delete pages you don’t need and rotate any sideways pages.
Export to TXT or Word instead of copying.
Remove repeated headers, footers and page numbers.
Fix hyphenated line breaks.
Rejoin broken lines inside paragraphs.
Clean invisible characters, spacing, ligatures and quotes.
Run a spell check and read the result once.
Most short extracts only need steps 4 and 6. Longer reports, contracts and research papers benefit from the full routine, and once you’ve done it a couple of times it takes minutes rather than an afternoon.
A clean copy of the text also makes everything that comes next easier, whether you’re quoting a source, feeding a document into a translation tool, building a spreadsheet from a report or pasting a policy into your website. Spend five minutes on the cleanup and you won’t spend an hour fixing the consequences later.