Compress PDF

What 'compress' actually does to a PDF

A PDF is a container of objects — fonts, images, vector paths and the text layer — wrapped in cross-reference tables. SnapToolz's compressor works on the container, not the contents: it removes orphaned objects, deduplicates repeated resources, packs many small objects into compressed object streams, and drops redundant metadata. Nothing visible changes because no image is resampled and no font is subset away. That's why the output is visually bit-identical and safe for contracts or print.

The trade-off is the ceiling. Structural compression typically recovers 10-40%, and a PDF that Acrobat already saved with object streams enabled has almost nothing left to give. The big wins come from PDFs exported by simpler tools, scanners, or 'Print to PDF' drivers that never optimised the container in the first place.

Why image-heavy PDFs are the real prize

If your PDF is mostly scanned pages or photos, the bytes live in the images — and structural compression leaves images untouched by design, so it won't move the needle much on a scan bundle. The honest fix is to compress the images first, then assemble: run the source images through Compress Image (or convert them to WebP with Convert Image), then build the PDF with JPG to PDF. A 50 MB scan bundle routinely drops below 8 MB this way with no readability loss.

For a born-digital PDF (text, vectors, a few logos), the structural pass is usually all you need to clear an email or portal limit.

Hitting a hard size target

Email gateways and government portals often impose exact ceilings — 100 KB, 200 KB, 2 MB, 5 MB. If a single pass doesn't get you under the line, the reliable strategy is split-and-rebuild: separate the document with Split PDF, drop pages you don't need, recompress the images, and merge the survivors back with Merge PDF. The dedicated walkthrough — compress a PDF to 100KB, 200KB or 500KB — covers the exact moves for the tightest targets.

Compression levels, decoded

Every level leaves text, fonts and vector graphics untouched and fully selectable — only embedded images are re-encoded. A PDF with no images therefore compresses very little at any level, and all three produce the same file: text is already stored compressed, so there is nothing left to squeeze.

Before you compress a document you'll send

Compression strips redundant data, but it is not redaction — text hidden under a graphic, or in a layer you can't see, survives the pass and stays extractable. If the file is confidential, redact and verify first (see privacy-cleaning a file before you share it), then compress. And keep the original: text survives untouched, but embedded images are re-encoded, so the master is still the file you want for future editing or printing.