webtools

Compress PDF

Private: nothing leaves your browser

Redrawn on your device

Drop PDFs here, or click to choose them

As many as you like at once. Up to 100 MB each. Best results on scans and documents full of pictures.

Choose a PDF to see what can come off it.

Choose a PDF to compress it.

Shrink a PDF that is too big to email by redrawing every page as a compressed image. It works very well on scans and photo-heavy documents, and it is the wrong tool for a page of text, which this page will tell you rather than quietly handing you a bigger file.

About compressing a PDF

Nearly all of the weight in a large PDF is pictures. A scanner producing one high resolution image per page is how a ten page document becomes forty megabytes, and no amount of clever restructuring fixes that, because the pictures are the file. This tool redraws each page at a sensible screen resolution and stores it as a JPEG, which is what actually removes the weight. A scanned contract routinely comes back at a fifth of its original size and is still perfectly readable.

Being plain about the cost: the result is pictures of your pages. Text in it can no longer be selected, searched or copied, links stop working, and form fields become part of the image. For a scan that was already just pictures you lose nothing, which is why this is the right tool for one. For a document that was typed, you are converting text into a photograph of text, and that usually makes the file larger, not smaller. If such a file is only awkward because of how many pages it has, splitting it gets you under a limit without touching the text at all.

That case is handled honestly rather than hidden. If the compressed version is not actually smaller, the page says so and gives you your original back untouched instead of pretending it did something. It all happens in your browser, so nothing is uploaded, and there is no queue, no size limit imposed by somebody's free tier and no copy of your document on anyone's server.

How to compress a scanned PDF

  1. Drop a PDF on the box, or click to choose one.
  2. Pick a strength. Balanced suits almost everything that will be read on a screen.
  3. Tick greyscale if the colour does not matter, for instance a signed black and white scan. It takes off a good deal more.
  4. Press Compress. The status line counts the pages as it works.
  5. Check the saving that is reported before you use the result. If it says the file got bigger, it did, and your original was returned instead.

Common questions about compressing a PDF

Does my PDF get sent to a server to be compressed?

No. The document is opened, redrawn and rebuilt entirely by JavaScript on this page, so it stays on your own device. Nothing is sent, logged or stored. That is worth more here than on most tools, because the PDFs people need to shrink are usually scans of something personal. You do not have to take that on trust. Open the Network tab of your browser's developer tools and compress a file: every request listed is for this page's own code, served from this domain, and not one of them carries your document. Compress once so that code is cached, then disconnect from the internet and compress another, which a site that uploaded your file could not do.

Will the text still be selectable afterwards?

No, and that is the honest trade at the heart of this tool. Every page becomes an image, so text cannot be selected, searched or copied, and links and form fields stop working. If your PDF came from a scanner it had none of those to begin with. If it was typed, keep the original as well, and consider whether you need to compress it at all.

Why did my PDF get bigger after compressing it?

Because it was a text document, and this tool works by turning pages into pictures. Stored as text, a page is a few thousand bytes of characters and font references. Stored as a picture of that text, it is hundreds of thousands of pixels, and no JPEG setting makes a photograph of a paragraph smaller than the paragraph itself. This page measures both and hands back your untouched original when the compressed version is the larger of the two, which is why you can see the comparison at all: a tool that only ever shows you a result would have given you the bigger file and called it compressed. There is usually nothing left to take off a file like that. A PDF exported from a word processor is already close to as small as its contents allow, so the real options are to remove pages, split it, or send it as it is.

How much smaller should I expect?

On a scanned document, roughly 40% off on Light, 60% on Balanced and 70% or more on Strong, and greyscale takes off more again. On a PDF exported from a word processor, expect no saving at all. The exact figure depends entirely on what is in the pages, which is why the result is measured and reported rather than promised in advance.

Can I keep the text and just shrink the images?

That is the ideal answer and it is not what this does. Re-encoding the embedded images while leaving the text alone means taking the document structure apart and rewriting it, which needs far more than a browser page can reasonably carry. Desktop tools such as Ghostscript do it. Restructuring a PDF losslessly, which some sites call compression, was tested here and saved nothing measurable on files from any modern producer, so it is not offered as a mode that appears to do something while doing nothing.

Next

Related tools

All PDF tools