Remove Punctuation & Characters

About the Remove Punctuation & Characters

Tick the kinds of characters to remove — punctuation, symbols, numbers, letters, emoji, spaces, line breaks, invisible and control characters or non-ASCII characters — or type your own list, such as #@* or a range like a-z. Switch to Keep only to do the opposite: keep, say, letters and spaces, and remove everything else.

Each kind follows the Unicode character categories, so it works in every language: punctuation includes curly quotes, dashes and the Hindi danda, numbers include ², ½ and Arabic digits, and an emoji with a skin tone or a family emoji is removed as one character. A report shows how many characters each rule removed and which characters were removed most often.

Common uses

• Cleaning text for word clouds, word counts, search and machine learning.
• Removing emoji and symbols from names, titles and product data before importing them.
• Finding and removing zero-width spaces, byte order marks and other invisible characters that break code, CSV files and searches.
• Keeping only digits from phone numbers, prices or order IDs.
• Making text ASCII-only for systems that reject other characters.

What each kind covers

Kinds follow the General Category of each character in the Unicode Character Database (UAX #44):

• Punctuation (P): . , ; : ! ? quotes, brackets, dashes, … and the punctuation of other scripts (। ، 。). In Unicode, # % & * @ are punctuation too.
• Symbols (S, without emoji): $ € ₹ + = < > | ~ ^ ` © ® ™ ° arrows and box drawing.
• Emoji: whole emoji sequences, including skin tones, flags, keycaps and ZWJ families.
• Numbers (N): 0–9, digits of other scripts (٣, ३), ², ½ and Roman numerals such as Ⅻ.
• Letters (L): letters of every script, removed together with their accents.
• Spaces and tabs: the normal space, no-break spaces and other space characters, and tabs. Line breaks: LF, CR LF, CR and the Unicode line and paragraph separators.
• Invisible and control characters: format characters such as zero-width spaces and joiners, the byte order mark and soft hyphens, control characters, private-use and unassigned code points. Joiners inside an emoji stay, so the emoji is not broken.
• Non-ASCII: anything outside the 128 ASCII characters, including accented letters and emoji.

Your own characters

Type the characters to match, one after another: #@* matches #, @ and *. A dash between two characters makes a range: a-z, 0-9, A-F. Use \n for a line break, \t for a tab, \s for a space, \- for a dash and \\ for a backslash. Tick Ignore capitals to match a-z and A-Z alike.

Replacing and tidying

Removed characters can be replaced with nothing, a space (so end.Start becomes end Start, not endStart) or your own text. Tidy up spaces then turns double spaces into one and removes spaces at the start and end of each line, which is usually what you want after removing punctuation; untick it to keep spacing exactly.

Limitations

• Works on up to 10 million characters at a time; split longer files into parts with the Text Splitter first.
• Removing zero-width non-joiners changes how some Persian and Indic words are drawn; untick Invisible and control for such text.
• Keeping ’ and - inside words works for apostrophes and hyphens between letters or digits, not for other punctuation.
• Your own characters are matched exactly; regular expressions are not supported (use Find and Replace for patterns).

Privacy

Everything happens in your browser. What you enter or open here is not uploaded or stored by MySmartCoPilot.