🤏 PDF Compressor

Shrink PDF file size locally — re-encodes embedded JPEG images at the chosen quality. Text stays selectable. Files never leave your browser.

Drop PDFs here to compress

Re-encodes embedded JPEG images; text and fonts are preserved

Click to browse, or drop anywhere — up to 25 files

FAQ

Does my PDF ever leave my computer?
No. All parsing, image re-encoding, and saving happen locally using pdf-lib. PDFs often carry sensitive content (contracts, IDs, financial records), and none of it is uploaded anywhere.
What does each preset actually do?
Each preset re-encodes every embedded JPEG image at a fixed quality: Lossless q=0.92, Visually identical q=0.85 (default), Smaller q=0.70, Smallest q=0.50. 'Lossless' is misleading shorthand — true bit-perfect PDF compression isn't browser-feasible — but at q=0.92 the visual loss is imperceptible.
Why is text still selectable after compression?
Only embedded image streams are re-encoded; text glyphs, fonts, page content streams, and annotations are untouched. The output is a normal PDF whose text is selectable, copyable, and searchable in any viewer.
Why didn't my PDF get smaller?
This v1 only re-encodes JPEG (DCTDecode) image streams. PDFs that are text-only, or whose images are stored as raw pixel data (FlateDecode), or that use the rarer JBIG2 / JPX / CCITT filters, will pass through unchanged. Photo-heavy PDFs typically shrink 30–70%.
How is this different from Acrobat or Ghostscript?
Desktop tools can subset fonts, downsample colour profiles, and use specialised PDF filters (JBIG2 for black-and-white scans, lossy JPX). This tool sticks to what's safe in a browser — JPEG image re-encoding plus structural save options — so it can't beat them on every file, but it never uploads anything.

More tools

JSON

Code

Ornaments

Image

Web / SEO

Generators

PDF