A PDF or InDesign export
Publishing tools and print pipelines sometimes map a rule or a quotation dash to U+2015. Copy a paragraph out of the PDF and the character comes with it.
Free · Private · No sign-up
U+2015 looks exactly like an em dash and behaves like nothing else. Paste the text and turn every one of them into a real em dash, a hyphen, or nothing.
The horizontal bar (―, U+2015) is a separate Unicode character from the em dash, and on screen the two are indistinguishable in most fonts. The difference is what software does with them. An em dash (U+2014) is the character editors, style guides and word processors mean by “dash”, and it is the one your keyboard shortcut, your autocorrect and your typography settings produce. The horizontal bar is the alternative a handful of typesetting traditions prefer for quoted dialogue, and the one several publishing pipelines emit without being asked.
Because they look identical, nobody notices one is in the text until something rejects it. A form validator with an ASCII or a small character whitelist refuses it. A style checker flags it as an inconsistency against the em dashes elsewhere in the same document. A find-and-replace for “—” reports no matches on text that visibly contains one. And in a font that never drew U+2015, the character renders as an empty rectangle — the box that sends people looking for “why is my em dash a box”.
This tool removes exactly that character and leaves genuine em dashes, en dashes and hyphens alone. The default replacement is a real em dash, which is what most people actually want: same appearance, same meaning, but a character that every tool agrees on. Choose Hyphen if the destination is a code field or a slug, or Nothing if the bar was decoration — a rule, a separator, an unwanted divider line — rather than punctuation.
Nobody types U+2015 on purpose. It arrives from a document, a pipeline or a font.
Publishing tools and print pipelines sometimes map a rule or a quotation dash to U+2015. Copy a paragraph out of the PDF and the character comes with it.
Some house styles open dialogue with a dash instead of quotation marks, and specify the horizontal bar for the job. If a manuscript follows that style, every one of those dashes is U+2015.
A document opened on a machine without the original font can have its em dashes rewritten by a fallback that uses the horizontal bar — the glyph looks right, the character does not.
Scrapers, OCR passes and older CMS platforms normalise punctuation with their own mapping tables, and some of those tables write U+2015 where the source had U+2014.
The right one depends on whether the bar was punctuation, a separator or a rule.
The right answer almost every time. Identical appearance, and every editor, validator and chatbot agrees on what it is.
ASCII and universally accepted. Use it when the text has to survive a spreadsheet, a URL or a database field.
When the character divided two labels rather than punctuating a sentence, a space is what the line actually wanted.
Delete it outright when the bar was a decorative rule — a horizontal line of dashes standing in for a section break.
Whichever you pick, it applies to the whole paste in one pass — a single chapter, a whole manuscript or a spreadsheet column. Nothing is uploaded and there is no word limit, so a page of cases is as safe as one sentence.
The horizontal bar is one of several dashes that are not the em dash you expected.
The character, its code point, and why it breaks validators.
Take out real em dashes instead of converting to them.
Fix the two- and three-em dashes that show as empty boxes.
Turn every dash in the text into one consistent style.
Take out U+2011 so pasted numbers search and import again.
Compare all 11 dash characters side by side with their code points.
No sign-up, no upload, no word limit. Paste, pick, and copy the result.
Remove dashes now