Edit PDF Text

Change printed text and redraw it with a built-in font

1Choose file2Settings3Process4Save

Choose a PDF and its contents will be analyzed and shown here.

Files are processed only in this browser. Nothing is uploaded to a server.

Pick text on a page, change it and redraw it with a built-in font

Continue in another tool
Task settings

The original text is only covered up; it stays inside the file. This tool is for fixing typos, dates and names. Don't use it to remove (redact) personal information. It can't reuse the original fonts and redraws with built-in fonts (Noto Sans or standard fonts), and it only edits within one line (no reflowing of long text).

Click a run of text in the preview or pick one from the list to open the edit panel. You can fix several spots on several pages in one go.

Remove the password from an encrypted PDF first. Processing focuses on page content, so bookmarks, attachments, forms and signatures may not be preserved. Keep the original and check the result.

About Edit PDF Text

In a word processor, changing one letter pushes the rest of the line along. A PDF has no such structure. Text on a page is stored as drawing instructions — put these glyphs, in this font, at these coordinates — and the idea of a sentence, a line or a paragraph is something the viewer reconstructs by looking at positions. Editing PDF text therefore comes down to hiding the old drawing and placing a new one on top.

That is exactly what this tool does. It reads the position and size of the text run you click (via PDF.js), paints a rectangle in the background colour over that box, and draws your new text with a bundled font. The original glyphs are not deleted from the file, so copying or searching the result may still turn up the old wording.

Why the original font can't simply be reused

To keep files small, PDF producers usually embed only the characters a document actually uses. This is called subsetting, and the PDF specification (ISO 32000-1, section 9.6.4) requires such fonts to carry a tag of six uppercase letters and a plus sign, as in ABCDEF+Garamond. If you see names like that in a viewer's font list, the fonts are subsets.

For English text the gap is often small — a contract will contain most of the alphabet — but not always. Replace “Smith” with “Szabo” in a document that never used a capital Z and the embedded font simply has no Z to draw. It gets much worse in Korean, Chinese or Japanese: Korean alone has 11,172 precomposed syllables, and a one-page letter might use a few hundred of them. On top of that, many PDFs store character codes as font-internal glyph IDs rather than Unicode, so even an existing glyph is awkward to address.

So the tool doesn't try. It draws with Noto Sans KR, Noto Sans JP or one of the standard PDF fonts (Helvetica, Helvetica Bold, Times, Courier). A sans-serif patch in a serif body is visible when you zoom in, which is why the tool suits short fixes like a date, an amount or a name.

What happens when a character isn't in the chosen font

The standard 14 fonts only cover WinAnsi, a Western European character set: accented Latin letters are fine, but Greek, Cyrillic, Korean or Japanese are not. If you pick Helvetica and type Hangul, the tool switches that single edit to Noto Sans KR automatically, and kana or kanji switch it to Noto Sans JP; the edit list shows the font that will really be used. The Noto Sans KR file on this site has no kana or kanji and Noto Sans JP has no Hangul, so when the chosen Noto font can't draw the text at all (Noto Sans KR with Japanese only, for example) the other one is used. Text mixing Hangul and kanji stays in the chosen font, so the kanji can go missing.

If some character can't be measured even with the Noto font — typically an emoji or an unusual symbol — the tool stops with an error saying the text contains characters it cannot draw, rather than silently leaving a blank. New fonts are embedded as subsets too, so a couple of Korean corrections add kilobytes, not the full multi-megabyte font file.

When the new text is longer than the box

Because a PDF has no line-wrapping information, longer text can't flow onto the next line. The tool shrinks the font size until the text fits, but never below half the original size. If it still overflows at half size, it draws anyway and shows a warning; the text may then run into a neighbouring column or table rule, and rewording is the better fix. Line breaks you type become spaces.

Many PDFs also split a single phrase into several runs, usually because of kerning or tracking adjustments. “Invoice total” might be stored as “Invoice to” and “tal”, so the run you click can be shorter than you expect. Edit each piece separately in that case.

Things cover-and-redraw can't do

  • Remove sensitive information. The old text data remains and can be copied or found by search. Use a true redaction tool that deletes the content.
  • Edit scanned PDFs. The letters are pixels, so there is no text run to click.
  • Blend into patterned or gradient backgrounds. The cover is a single flat colour. You can switch the cover off and draw over the old glyphs, but that usually looks worse.
  • Keep a digital signature valid. Any change to the content breaks signature validation.

Worked examples

Correcting a date on a quote

Input
Original run “12 March 2025” at 11 pt, new text “2 April 2025”, font Helvetica
Output
Drawn at 11 pt in Helvetica; fits the box without shrinking

Everything is in WinAnsi, so no font switch happens. The cover colour is sampled from the pixels around the box, so a light grey table cell is covered in that grey rather than white.

A longer name in a narrow cell

Input
Box 60 pt wide, new text measures 80 pt at 10 pt
Output
Redrawn at 10 × 60 ÷ 80 = 7.5 pt

Text width scales linearly with size, so the fit is a direct ratio. At 120 pt the required size would be exactly 5 pt, the half-size floor; anything longer is drawn at 5 pt with an overflow warning.

Hangul typed with a Latin font selected

Input
Font Times, new text “서울특별시”
Output
That edit is drawn in Noto Sans KR; other edits keep Times

Times cannot encode Hangul, so the tool resolves the font per edit instead of failing the whole document.

How to use it

  1. Open the PDF and click the text you want to change in the preview.
  2. Type the replacement and check size, font, text colour and cover colour.
  3. Look for overlaps or clipping in the preview, then save the PDF.

Frequently asked questions

Why can I still find the old text when I search the edited PDF?

The original glyphs are covered, not deleted. That is fine for corrections, but it means this tool must not be used to hide information.

Can I match a serif body font?

Times is available for Latin text. There is no Korean or Japanese serif option, and the document's own font can't be reused because it is usually a subset.

Can I edit several places at once?

Yes. Queue edits across pages and save once. Each bundled font is embedded only once per document.

Will this work on a scanned contract?

No. Scanned text is an image. You need an image editor, or OCR followed by re-creating the page.

Does editing break a digital signature?

Yes. Any content change invalidates it; the signer needs to sign the corrected version.

Sources

Last reviewed:

Privacy and how data is handled