How to Compress a PDF Without Uploading It Anywhere
Re-compress the images inside a PDF to shrink it — offline, in your browser, with nothing sent to a server.
You export a scanned contract, go to attach it, and the email bounces back: 38 MB, over the limit. Now you’re hunting for a way to make it smaller, and every result on the first page wants you to upload the contract first. To a stranger’s server. The one thing you’d rather not do with a signed document.
The Compress PDF tool skips that part. The file is read into memory, its images are re-encoded, and a lighter PDF comes back out — all in the browser tab you already have open.
Why PDFs balloon in the first place
Text is tiny. A page of prose is a few kilobytes. What blows up a PDF is almost always pictures: phone snapshots dropped into a report, a 300-DPI scan of a ten-page lease, product photos in a catalog. A camera these days spits out images that are 4000 pixels wide, and a PDF will happily embed every one of those pixels even when the page only ever gets viewed at a fraction of the size.
So the fix isn’t magic. It’s the same thing your camera roll does when you share a photo: knock the resolution down a bit, re-save the JPEG at a lower quality, move on. Do that to every image in the file and a 38 MB monster often lands somewhere under 10.
What the tool changes, and what it leaves alone
Two dials, and that’s on purpose:
- Image quality — the JPEG quality of each re-encoded photo. Push it low and the file shrinks hard, but you’ll start to see the blocky softening JPEG is known for. Around 70% is a decent starting point for scans.
- Maximum image size — the longest side any photo is allowed to keep. Cap it at 1600 pixels and a 4000-pixel scan gets scaled down before it’s re-saved, which is usually where the biggest savings hide.
Everything that isn’t a photo gets copied through byte-for-byte. Your text stays text — selectable, searchable, sharp at any zoom. Vector logos and line drawings don’t get touched, because rasterizing them would make them worse, not smaller.
When it won’t do much
Here’s the honest part. If your PDF is mostly text and tables — an invoice, an ebook, a code printout — there are no fat images to squeeze, so the tool rewrites the file and tells you it was already well optimized. It will never hand you back something bigger than what you fed it. If nothing shrinks, you’ve lost nothing but a couple of seconds.
One trade-off worth knowing: bookmarks, the document outline, and interactive form fields live at the whole-document level, and they don’t survive the rebuild. The pages, their text, and their images all do. If you’re compressing a fillable form you still need to fill in, keep the original around.
Nothing leaves the machine
The reason upload-based compressors exist at all is server economics — someone has to pay for the CPU that re-encodes your images, and they recoup it with ads, subscriptions, or a vaguer plan involving your documents. A tool that runs the re-encode on your own device sidesteps the whole arrangement. There’s no file on someone else’s disk to leak, and you can watch it work with the network cable pulled.
Drop a heavy PDF into the Compress PDF tool, nudge the quality until the size looks right, and download — the file never leaves your machine.