Edit Text in a PDF Online: Fix It Without the Source File
Click a line of text in a PDF, change it and download the result. Nothing is uploaded and no sign-in is needed. Only lines of text that already exist can be edited.

Edit PDF changes text that is already in a PDF, directly in your browser, one line at a time. Nothing is uploaded, there is no sign-in and nothing to install, which suits the case where all you have is an exported PDF and no Word file or design source. It cannot add a new line, rewrite a paragraph or wrap text automatically, and it cannot edit the text in a scan.
How to edit: open a file, change a line, download
Drag a PDF onto the page or click Choose file. A password-protected file asks for its password first. Files up to 32 MB are accepted.
Move the pointer over a piece of text and a box appears. Click it, type the new text, and press Enter to confirm or Esc to discard. Use the Undo button or Ctrl+Z (⌘Z on a Mac) to step back, all the way to the file as you opened it. Clear a line and confirm, and the line is deleted.
When you are done, click Download PDF or press Ctrl+S (⌘S on a Mac). Until you download, your changes exist only in this tab: the status bar reads "Unsaved changes" and the browser warns you before the page closes. The original file on your computer is not overwritten; you download a new one.
Your file stays in your browser
Reading, editing and producing the new file all happen in the browser. The PDF is not uploaded, so a contract draft or a form with personal data never leaves your computer. One kind of request does go to PDF123: when you type a character that the file's own fonts do not contain, that character alone, with no file and none of the rest of what you typed, is sent to fetch a font for it. See "Which font the edited text uses" below.
What it is good for
Small edits of about the same length, where the line stays as long as it was and the content around it stays in place. On an invoice, that means changing the total from ¥1,250.00 to ¥1,520.00, changing the invoice date to October 8, 2026, and correcting a typo in the buyer's name, with everything else left as it was. A unit price on a quote, a phone number on a résumé, a date or a name in a notice are the same kind of edit.
Which font the edited text uses
Characters that the line's original font already has keep that font. Characters it lacks get a similar one.
Many PDFs store only the glyphs they actually use. Changing one digit in an amount to another digit therefore usually shows no difference, while a character that never appeared in the file, say a rare Chinese character, has to be drawn with a different font and may look slightly unlike its neighbours. In that case PDF123 fetches the few matching glyphs from an open-source font and embeds them in your file. Chinese characters follow your interface language: Simplified forms in the Simplified Chinese interface, Traditional forms in the Traditional Chinese interface, and Japanese or Korean forms in the Japanese and Korean interfaces.
While the font loads, the toolbar shows "Loading a matching font…". If it cannot be fetched you get a notice and a built-in font is used; the text itself is still correct. For why a file can lack a character in the first place, see Why Font Subsetting Breaks Editing But Not Reading.
Is the old text really gone
Yes, but only the text you changed. The original text is removed from the page content, so copying or searching no longer finds it.
If the same text also appears elsewhere, for example in an annotation or in the document properties (title, subject), it does not change with it. To remove old text completely, use the Redact tool; see Before You Share: What Sanitize and Redact Actually Do.
What it cannot do
- It cannot add text. You cannot type into blank space or insert a new line. To put a line of Latin-script text on a page, such as a mark, use Add Stamp, which supports Latin characters only.
- It does not wrap text. Editing works on one line at a time: a line made longer keeps running to the right and whatever passes the page edge cannot be seen, and a line made shorter leaves a gap.
- It cannot edit the text in a scan or in an image. A scanned page is a picture with no text to click; if a box appears when you hover, it is real text. OCR (optical character recognition) can read the words in a scan, but it returns Markdown, not an editable PDF.
- It edits text only. Images, vector graphics and form fields cannot be changed.
- It does not support Arabic, Hebrew or Indic scripts (Devanagari, Bengali, Tamil and others). These need right-to-left or joined layout: typing them is refused with a message, and right-to-left lines cannot be opened for editing.
- Thai can be typed and the characters are right, but the tone marks and vowels stacked above and below a letter may sit slightly off.
- After you edit a file with a digital signature, the original signature is no longer valid.
When to go back to the source file
When a change moves what follows it, such as rewriting a paragraph or adding rows to a table, go back to the source the PDF was made from and export it again. If you only have the PDF, you can convert it to Word first and edit there. For a scan, read the text with OCR first.
Keep reading
- Why Font Subsetting Breaks Editing But Not Reading: why some edited characters end up in a different font.
- How OCR Reads a Scanned PDF — and Why This Tool Returns Markdown: what it can and cannot read when all you have is a scan.
- Before You Share: What Sanitize and Redact Actually Do: how to clear old text that still sits in annotations or properties.