PDF tool
Compress PDF
Make a PDF smaller. There are two ways to do that and they are not equivalent, so this tool makes you choose, and tells you what each one destroys.
Drop a PDF here
Or press to choose one.
nothing is uploaded — this page does the work
Why there are two buttons instead of one slider
Almost every free PDF compressor online has a single control and does not tell you which of these two things it is doing. Most of them quietly do the second one. You upload a 40 MB report, you get back a 3 MB file, you are delighted, and four months later you try to copy a paragraph out of it and discover that your document is now a stack of photographs. Nobody told you, and by then the original is gone.
So this tool makes you pick, names both options for what they are, and puts a warning in front of the destructive one. That is a worse conversion funnel and a better tool.
Restructure: what it can and cannot save
A PDF accumulates junk. Editing software leaves behind objects nothing references any more, writes the cross-reference table in its longest form, and stores small objects individually where they could be packed together. Re-saving cleans all of that up. Text, fonts, images and layout are untouched — the output renders identically and every word is still selectable.
On a file exported straight from a word processor, expect two to twenty percent. On a file that has been edited and re-saved several times, sometimes much more. On a file that is already optimised — anything produced by a modern PDF library, including one produced by this tool — expect nothing, and this tool will tell you it found nothing rather than handing back a file the same size and calling it compressed.
It cannot recompress images. If your 40 MB PDF is 40 MB because it contains twelve photographs at full camera resolution, restructuring will save you almost nothing, because the photographs are already JPEGs and are the entire file. That is the case where the other mode exists.
Rasterise: the right tool for exactly one job
Each page is rendered to pixels at the resolution you choose, saved as a JPEG, and placed into a new PDF at the original page dimensions. The output prints at the correct physical size; it simply contains a picture of a page rather than a page.
This is the correct thing to do to a scan. A scanned document is already a picture, often stored at 300 or 600 DPI when 150 would be indistinguishable in a document nobody will reprint. Rasterising it at 110 or 150 DPI routinely takes 90% off the size and costs you nothing you had, because there was never any real text in it to lose.
It is the wrong thing to do to a document that was born digital. You would be throwing away crisp vector text and replacing it with a photograph of crisp vector text, which is both larger than it needs to be at any readable resolution and unsearchable.
Resolution is the dial that matters, more than JPEG quality. 72 DPI is fine on a screen and unreadable printed. 150 DPI prints acceptably. Above 200 you are usually making the file bigger for no visible gain.
Why this one can take a moment
Restructuring is near-instant. Rasterising renders every page, and rendering is real work: a hundred-page document at 150 DPI is a hundred images at roughly 1275 by 1650 pixels each, drawn by your own device. On a laptop that is a few seconds. On an older phone it can be a minute, and the progress bar is honest about where it has got to.
It still happens entirely in this page. Nothing is uploaded either way, which also means there is no file size limit imposed by us — only by how much memory your browser is willing to give this tab.