The horizontal bar (U+2015)
Typesetting exports use it where you would type an em dash. Same length, different character, different behaviour in every parser.
Free · Private · Runs in your browser
PDF em dash remover — clean the long dashes out of a PDF
Upload the document, pick the replacement, press once. Every em dash — and the look-alikes PDFs actually emit — is rewritten in the extracted text, ready to copy or download.
Clean a PDF in your browser
Text with no long dashes
The text tools on this site run on pasted text, but most dash-riddled documents live as PDFs — a report where every aside arrived as —, a thesis whose citations use the multi-em dashes, a manuscript where the typesetter's long dashes survived the export. This page closes that gap for the em-dash family: drop the PDF in, and the extracted text comes back with every U+2014, U+2015, U+2E3A and U+2E3B rewritten — and nothing else touched.
That last part is the point of a focused tool. PDF exports lean on the horizontal bar (U+2015) where you would have typed an em dash, designed documents carry two- and three-em dashes (U+2E3A, U+2E3B) from citation styles, and all of it arrives pre-mixed with the em dashes that were genuinely there. One pass rewrites the whole family — en dashes, hyphens and every other character pass through untouched.
The output is text you can use immediately. The cleaned document appears in a text box on the page, with a copy button and a .txt download beside it — paste it into Word, Google Docs or your CMS, or keep the file.
Everything happens in your browser. The file is read locally, the text is cleaned in memory, and the result is shown on the page — there is no upload step, no server, no copy retained anywhere. One honest limit: the tool reads the text layer, so it works on text-based PDFs — scanned pages carry their words as pictures, and there is nothing to extract. Password-protected files are refused, and the size limit is 25.0 MB per file.
Why PDFs dash so strangely
The characters behind “I copied it exactly and it still looks wrong”.
Multi-em dashes (U+2E3A/B)
Bibliographies and redacted names use them. Most fonts cannot draw them, so they arrive as boxes or not at all.
One family, one pass
The whole em-dash family is rewritten together, so a document that mixes all three variants comes out consistent — no stragglers to hunt down by hand.
Related dash tools
PDF pastes rarely stop at one broken character. The rest of the fix lives here.
PDF Dash Remover
All three dash families at once, from the same upload.
PDF En Dash Remover
The same fix for ranges and figure dashes from a PDF.
PDF Hyphen Remover
Un-hyphenate the words a PDF broke at line ends.
Soft Hyphen Remover
The invisible hyphens inside PDF words, gone in a click.
Horizontal Bar Remover
Only U+2015 — for text where the bar is the sole problem.
Line Break Remover
Join the lines the PDF broke at every printed row.
Questions
Which characters does this tool remove?
The em-dash family: the em dash (U+2014), the horizontal bar (U+2015) and the two- and three-em dashes (U+2E3A, U+2E3B). En dashes, hyphens and every other character are untouched — that is what makes the tool safe to run on mixed text.
What should I replace PDF em dashes with?
A space for prose (the sentence stays readable), a hyphen when the dash marked a range or compound, or nothing when the dash was decoration. The chips above cover all three; the custom field takes any string.
Why does my PDF text show boxes instead of dashes?
Because the PDF used a character the receiving font cannot draw — usually a multi-em dash from a citation. This tool replaces those with a character you choose, which ends the box problem immediately.
Is my PDF uploaded to a server?
No. The file is processed entirely in your browser: pdf.js extracts the text locally, the dash engine runs as ordinary page JavaScript, and the cleaned text is written on your device. There is no upload endpoint to send anything to.
Does it work on scanned PDFs?
No. A scan is a picture of text with no characters underneath, so there is nothing to extract. If the scan reports zero dashes on a document you know is full of them, that is the tell. Run the pages through OCR first, then clean the resulting text-based PDF.
What about password-protected files?
They are refused with a clear message. Remove the protection — or print the document to a new PDF — and the copy will process normally.
How large a file can I clean?
25.0 MB per file. Larger documents are usually fine split in half — PDF readers can extract a page range into a smaller file, and each half cleans independently.
Ready to clean your PDF?
Drop the file in, pick the replacement, press once. Nothing is uploaded, nothing is stored — the clean text is yours in seconds.
Clean a PDF now