How to compress a PDF without losing quality
'Compress' and 'without losing quality' are in tension — but not as much as you might think. A guide to shrinking PDFs sensibly, from 100 MB monsters to email-ready.
Where PDF weight comes from
Open the file and check what's inside. PDFs are big for exactly three reasons:
- Embedded images at absurd resolution. A scanned page at 600 DPI weighs ~5 MB. A 20-page scanned contract is ~100 MB before it does anything else.
- Fonts embedded multiple times. Poorly-written export tools embed the full font file for every subset used, then never deduplicate. This can add 2–20 MB.
- Unused objects and revision history. PDFs are append-only by design — every "save" can leave old versions of every object behind. A PDF edited 10 times can carry 90% garbage.
The right compression approach depends on which of these is dominant.
The safe first move: recompress images
For most PDFs, 90%+ of the size is images. Reducing image DPI from 600 → 150 typically shrinks the file by 70–90% and looks identical on-screen. Print quality drops a bit at 150 DPI but is still perfectly acceptable for most uses.
Compress PDF lets you pick a quality preset. Try "medium" first — it usually cuts file size in half without any visible difference. Only step down to "small" if you specifically need to hit a size cap.
When you need to hit a specific size
Email caps at 25 MB, some forms cap at 5 MB, some at 2 MB. Compress to Target Size works backwards from the target — you tell it "make this 10 MB" and it picks the settings. This is much less frustrating than repeatedly guessing.
When quality loss is not acceptable
If the PDF is going to be printed on a professional press, signed and legally binding, or archived for compliance, don't recompress images. Instead:
- Strip unused objects. Not every tool exposes this, but the effect is often 20–40% size reduction with zero quality change. Some browser tools quietly do it during any re-save.
- Split and re-merge. Split PDF into single pages then Merge PDF back together. This rebuilds the object table from scratch, dropping accumulated garbage. It's crude but surprisingly effective on old, heavily-edited files.
- Convert scanned pages to text. If your "PDF" is really a stack of page images, running OCR and re-typesetting is the only path to a genuinely small file. This is beyond one-click tools.
What "lossless" actually means for PDFs
Lossless compression exists in PDFs too, but only for a narrow category: PDFs that are already text-based, generated from Word or a browser rather than scanned. On these, tools can save 10–30% by compressing streams more aggressively — the visible content is bit-for-bit identical.
For scan-based PDFs, "lossless" is not really possible in a compression tool. You are trading detail for size no matter what; the question is whether the trade is visible.
A practical playbook
- Under 5 MB target: accept some quality loss, use Compress to Target Size.
- Under 25 MB (email): Compress PDF medium preset almost always gets you there.
- Archival, no quality loss: don't compress. Split/merge instead if you need to strip garbage.
- The file is enormous because it's 500 pages: don't compress — Split PDF into chunks people can actually read.
Do it locally
Server-side compressors get files under limits by aggressive re-encoding, but you're uploading the original PDF to their servers first. On a signed contract or a medical form that's an awkward trade. Compress PDF runs in your browser — your file never leaves your machine.