CHUBBY PDF.
Your file never leaves this tab.
Everything runs in JavaScript (and WebAssembly) on your device. We can't see your PDF, and neither can our servers — because your file never touches them.
HOW COMPRESS PDF WORKS.
01. Drop
Drag your file onto the box, or click to pick one from your device.
02. Tweak
Set the options in the sidebar. Defaults work for most.
03. Download
Hit the big button. Your file is ready in seconds — never leaves your browser.
ABOUT COMPRESS PDF.
PDF file size ballooning out of control is almost always an images problem, not a text problem — a handful of full-resolution scanned pages or embedded high-DPI photos can turn a 20-page document into a 50MB monster that bounces off an email attachment limit. minipdftoosz's Compress tool tackles this directly: it walks through every embedded image in the PDF, re-encodes it at a lower JPEG quality and, for the more aggressive settings, downsamples its resolution, then rebuilds the PDF around the smaller images using pdf-lib.
Three compression levels are offered because "smaller" and "same quality" genuinely trade off against each other. Light compression barely touches image quality and mostly just cleans up redundant PDF object streams — a safe default when you're not sure how a document will be printed later. Medium is the recommended middle ground for everyday sharing. Heavy pushes image quality down further and is best reserved for documents that are mostly text with a few incidental images, where nobody will zoom in on a photo anyway.
Because this runs in JavaScript inside your tab rather than on a remote compression server, there's no size cap imposed by an upload quota, and — just as importantly — no copy of your (possibly sensitive) scanned document sitting in a compression queue on someone else's infrastructure while it processes. The trade-off is that very large files (many hundreds of megabytes) will take longer on a modest device than the same job would on a beefy server, since your own CPU is doing the work.
Compression is not lossless for images — heavy settings visibly soften photos if you zoom to 400%. It is effectively lossless for the surrounding PDF structure: text stays as real, selectable text (not flattened to an image), form fields keep working, and page count never changes. If a document is already mostly plain text with no images, don't expect compression to shrink it much; there's nothing left to squeeze.
It's worth running Compress as close to the last step in your workflow as practical — if you still plan to merge, watermark, or add pages afterward, compress after those steps rather than before, so the extra content you add doesn't reintroduce size you already trimmed. For documents you'll compress repeatedly (a recurring report template, for instance), keeping a note of which compression level looked acceptable last time saves a bit of trial and error on future runs.
Anyone emailing scanned documents runs into size limits eventually, which is where Compress earns its keep most often — a mortgage application packet, a set of scanned IDs, a multi-page signed agreement that started as a phone-camera scan and ballooned past what a typical inbox will accept as an attachment. It's also useful proactively: compressing archival documents before long-term storage keeps a personal or business document library from quietly consuming far more disk space than the actual content warrants.
- Try Medium first — it catches most of the available savings without visible quality loss on typical scanned documents.
- If a compressed file still looks too large, check whether it contains embedded video or audio attachments, which this tool does not re-encode.
- Compress after merging, not before — merging doesn't add overhead, but running compression on each piece separately wastes the pass.
FAQ.
Common questions about compress pdf. Something else on your mind? Ask us.
Depends on the PDF. Text-only PDFs shrink less; image-heavy ones can drop significantly by re-encoding embedded images.
Only if you pick "heavy" — that re-encodes images at lower quality/resolution. Medium is safe for most documents.
No hard cap, but very large PDFs (500MB+) may be slow to process fully in-browser.
Image re-encoding (JPEG quality reduction / downscaling) and PDF object stream optimization via pdf-lib, done locally.