Fix Smart Quotes in Python Code

You've copied a Python snippet from a blog post, Slack message, Word document, or email. It looks perfectly fine. You paste it into your editor, run it, and get:

  File "script.py", line 3
    print(“Hello, world!”)
          ^
SyntaxError: invalid character '“' (U+201C)

The culprit is smart quotes (also called curly quotes or typographic quotes). These are Unicode characters that word processors, email clients, and messaging apps automatically substitute for straight ASCII quotes.

Before vs. After: what your code actually looks like

Here's the problem visualized. This is what you think you pasted:

# This looks correct but won't run
message = "Hello, world!"
name = 'Alice'
print(f"{name} says: {message}")

But this is what your editor actually received (Unicode code points shown):

# Hidden curly quotes — causes SyntaxError
message = “Hello, world!”       # U+201C and U+201D
name = ‘Alice’                  # U+2018 and U+2019
print(f“{name} says: {message}”)

After cleaning with Unformat (Developer mode), you get valid Python:

# Clean ASCII quotes — runs correctly
message = "Hello, world!"
name = 'Alice'
print(f"{name} says: {message}")

Python expects ASCII quotes

Python's lexer only recognizes two quote characters for string delimiters:

• Double quote: " (U+0022)
• Single quote: ' (U+0027)

But word processors and email clients silently replace these with:

• Left double quote: “ (U+201C) — looks like " in most fonts
• Right double quote: ” (U+201D)
• Left single quote: ‘ (U+2018) — looks like ' in most fonts
• Right single quote: ’ (U+2019)

These look nearly identical in most fonts, making the bug extremely hard to spot visually. The Python interpreter sees completely different characters and rejects them immediately.

Where smart quotes come from

Microsoft Word and Google Docs have “smart quotes” enabled by default. Any code snippet typed or pasted into these editors gets its quotes silently replaced.

Slack, Teams, and Discord apply the same transformation outside code blocks. When a colleague pastes code into a Slack message without wrapping it in backticks, the quotes are converted to Unicode curly quotes.

Blog posts and tutorials copied from web pages frequently carry smart quotes, especially if the content was authored in a CMS like WordPress or Medium that applies typographic transformations. Even Stack Overflow answers can contain them if the author composed their answer in a word processor first.

PDF documents almost always use smart quotes. Copying code from a textbook, research paper, or documentation PDF is one of the most common sources of this issue.

The problem goes beyond quotes

Smart quotes rarely travel alone. The same sources that produce curly quotes also inject:

• Non-breaking spaces (U+00A0) that look like regular spaces but cause Python's IndentationError: unexpected indent
• Zero-width spaces (U+200B) that create invisible characters inside variable names, breaking NameError: name is not defined
• Em-dashes (—) that replace minus signs and hyphens, breaking arithmetic expressions
• BOM markers (U+FEFF) at the start of text that cause SyntaxError on the first line

Unformat's Developer mode catches all of these simultaneously.

Use Developer mode for code cleaning. It replaces all Unicode quote variants with their ASCII equivalents and simultaneously fixes every other invisible character that breaks Python.

Developer mode handles the full set of problematic characters:

• “ ” → " and ‘ ’ → ' (smart quotes → straight ASCII quotes)
• U+00A0 → regular space (fixes IndentationError)
• U+200B U+200C U+200D U+FEFF → removed entirely (fixes phantom characters)
• — – → - (em/en-dashes → regular hyphens)
• Tab characters → spaces (configurable: 2 or 4 spaces, or keep tabs)
• BOM markers → removed (fixes first-line SyntaxError)
• \r\n → \n (CRLF → LF line endings)

Fixing it programmatically in Python

If you need to fix smart quotes in a Python script rather than manually, here's a quick approach:

def fix_smart_quotes(text: str) -> str:
    replacements = {
        '\u201c': '"', '\u201d': '"',  # double
        '\u2018': "'", '\u2019': "'",  # single
        '\u00a0': ' ',                  # nbsp
        '\u2014': '-', '\u2013': '-',  # dashes
    }
    for old, new in replacements.items():
        text = text.replace(old, new)
    return text

But when you just need to quickly clean a pasted snippet, Unformat is faster than writing a script. Paste, clean, copy — done in under a second.

All processing runs in your browser. Your code never touches a server, which matters when you're working with proprietary code or sensitive data.