Remove Watermark from PDF

It covers the watermark. It does not delete it.

The important part, and the reason this page exists.

A PDF is not a picture — it is a list of objects. Text is text, graphics are graphics, and each sits at a position on the page. A watermark is one of those objects.

This tool works by drawing a rectangle over the watermark, filled with a colour sampled from the surrounding page so the cover blends in. Visually the result is clean and the watermark is gone from view.

But the watermark object is still in the file. It has something on top of it, which is not the same as being removed. So anyone who copies the text out of the PDF, runs an extraction tool over it, or opens it in software that lists objects will still find the watermark text sitting there.

What that means in practice:

If you want the watermark out of sight — you are printing the document, reading it, showing it on screen, or sending it to someone who will simply look at it — this works and the result looks right.

If you need the watermark genuinely gone from the file — because the document is going somewhere it will be processed, indexed, searched or inspected — this is not enough, and the next section gives you a method that is.

How to remove it for real

There is a way to do this properly using tools on this site, and it takes two steps.

Flatten the document to images, then rebuild it. Use PDF to JPG to turn every page into a picture, then JPG to PDF to assemble those pictures back into a PDF. Because each page becomes a flat image, every object on it — including the watermark and anything underneath your cover — stops existing as an object. There is nothing left to extract.

Do the covering here first, then flatten. That way the cover is baked into the image and the watermark is genuinely gone.

What flattening costs, so you can decide: the text in the whole document stops being selectable and searchable, a screen reader can no longer read it, links stop working, and the file usually gets larger. It is a real trade, and it is the same trade every "permanent" solution makes.

The better option, if it is available to you: ask whoever produced the document for a copy without the watermark. A watermark is usually applied on export, so an unwatermarked version often exists and is one email away. For a licensed document, that is what the licence entitles you to.

Auto-detect is worth trying first

The genuinely clever part of this tool.

Auto-detect looks for two things that together describe almost every watermark: text set at an angle, and text that repeats on most pages. Ordinary body text is horizontal and different on every page, so the heuristic separates a watermark from the document remarkably well.

It works best on exactly the standard case — a diagonal "DRAFT", "CONFIDENTIAL", "SAMPLE" or company name stamped across every page. It will not find an image watermark, a logo, or a mark that appears on only one page, since neither fits the pattern.

So: run auto-detect, look at what it found, then fix up by hand. That is usually much faster than selecting everything manually, and it often catches a mark on page forty that you would not have scrolled to.

Selecting by hand, and applying across pages

Drag on the page preview to mark an area, and use Prev and Next to move through the document. Each selection appears in a list you can remove items from, so a mistake costs one click.

The setting that saves the most time is where your selections apply:

Every page, same spots is the right choice for a standard watermark, because it sits in the same place on every page. Mark it once, apply everywhere, done — this is the difference between a ten-second job and a three-hundred-page one.

The page they were drawn on is for marks that vary — a first-page stamp, a footer that only appears on some pages, or a document where the watermark moves.

One practical warning about apply-to-all: check a few pages afterwards, including the first and last. If a page has a different layout — a title page, a chapter opener, a landscape table — the cover may land on top of real content. That is the one way this tool can damage a document rather than just failing to fix it.

When the cover is visible

Two things to expect, both fixable.

A patterned or coloured background shows the seam. The cover is a single sampled colour, so it disappears on plain white and can show as a faint block on a textured background, a gradient, or a page with a coloured panel. Tighter rectangles help, and several small covers beat one large one.

Text underneath gets covered too. A watermark crossing a paragraph cannot be covered without covering the words it crosses, because the cover is an opaque rectangle. There is no way round that with this method — and it is the case where flattening plus a proper image-based repair, or simply getting a clean copy, is the only real answer.

Nothing is uploaded — the PDF is opened and rewritten in your browser, which matters since watermarked documents are usually somebody's confidential or licensed material.

Related tools

Remove Watermark from Image does this for a photograph, and it can promise the pixels are genuinely replaced, because an image has nothing underneath. PDF to JPG is the first half of the flattening method above. Extract Text from PDF is how you check whether the watermark text is still in your file — if it appears there, it is.