ALL PDF TOOLS37
Annotate PDF
Annotate PDF — Free Online Tool
Place highlight blocks, outline boxes and text notes on one page, then download the marked PDF.
About Annotate PDF
Annotate PDF puts three kinds of mark on a page you choose: a translucent highlight block, an outlined box, and a short text note. The page is rendered with PDF.js exactly as a reader will see it, and you place marks on that picture — drag to set a rectangle, click to move one, or type its position and size into the X, Y, Width and Height fields. Those fields are the same four numbers the drag writes, so the page and the rail can never disagree and nothing here needs a mouse. Two things this page cannot do, and they are the two people assume it can. It does not select text: there is no text layer under the preview and the engine has no model of where the words sit, so a highlight is a coloured rectangle at the spot you chose, not a highlight that finds a line and follows it across a column break. And the marks are painted into the page's own content rather than attached as PDF annotation objects, so a reader shows no comment to click, no author, no reply and no delete. That makes them permanent in normal use — taking one out again means editing the page. If you want something a colleague can answer, this is the wrong tool. The rest is detail worth knowing before you rely on it. A highlight is a flat translucent block laid over the page, not a multiply blend, so the text underneath comes through tinted, and at full opacity it covers that text completely. Notes are drawn in Helvetica and can only carry the characters the standard PDF fonts cover, which is Latin letters, digits and punctuation including accents, curly quotes and the euro sign; Chinese, Cyrillic, Greek, Arabic and emoji are refused before you press the button rather than after. A note is never wrapped for you, so press Enter where a line should break, and be aware that extra lines land twenty-four points below the one before whatever size you pick. Only the page you marked is changed, and only by adding to it. Page sizes and page order are untouched, existing text stays selectable, and no image is re-encoded.
Questions
Does this highlight text, or just draw a rectangle over it?
It draws a rectangle. There is no text layer under the preview and the engine has no model of where the words sit on the page, so a highlight is a translucent block you position over a line by eye. It cannot find a sentence, follow it across a line break, or stop neatly at the end of a word — you say where it goes, and it goes exactly there.
Can a reader delete these marks, or reply to them?
No. The marks are drawn into the page's own content rather than added as PDF annotation objects, so a reader shows no comment to click, no author, no reply and no Delete. In normal use they are permanent: taking one out again means editing the page content itself. If you need marks a colleague can answer or tick off, use a reader that writes real annotations.
Do marks land correctly on a page the file stores rotated?
Yes, and it is worth saying because it was briefly not true. A PDF page can carry its own rotation of 90, 180 or 270 degrees, and the engine used to shift a rectangle down along the unrotated axis — which only means "down the page" when the rotation is zero. A mark could land a full mark-height away, twice that at 180 degrees, and sideways as well on a quarter turn. The correction now follows the page as a reader sees it, and a test places a mark at the same spot on all four rotations and fails if any of them drifts. Notes were never affected: text is drawn about its own origin with no correction to get wrong.
Can a note be written in Chinese, Russian or emoji?
No. Notes are drawn with the standard PDF fonts, which only encode the Windows-1252 set: Latin letters, digits and punctuation, plus accented letters, curly quotes, the euro sign and similar. Chinese, Japanese, Cyrillic, Greek, Arabic and emoji cannot be drawn, and neither can a few Central and Eastern European letters such as ł, ș and ő. The page checks as you type and names the characters, so you find out before you press the button rather than after.
Can I mark more than one page in a single pass?
Not in this version. Every mark lands on the page shown in the preview, and switching pages carries the marks with you rather than leaving them behind. To mark a second page, download the result and drop that file back in — by then the first page's marks are part of the page, so nothing is lost.
Is my file uploaded to a server?
No. Transmute processes everything locally in your browser using JavaScript and WebAssembly. Your files never leave your device — there is no server, no upload, no cloud processing.