RunToolRunRunToolRun
โ† Blog

How PDF Compression Works โ€” and How to Shrink Files Without Ruining Them

July 20, 2026 ยท 2 min read

PDFs that will not fit in an email attachment almost always bloat for the same reason: the images inside them. Text contributes almost nothing to a PDF's size. A 300-page text-only PDF can be a few hundred kilobytes, while a 10-page scanned document can be 40 MB.

The four things that inflate a PDF

SourceTypical shareReducible?
Embedded images70-95%Yes โ€” this is where the win is
Embedded fonts2-10%Partly (font subsetting)
Metadata / revision history1-5%Yes, completely harmlessly
The actual text content1-3%No meaningful gain

The practical consequence: PDF compression is almost entirely image compression. When a tool says it shrank your file by 80%, what it did was re-encode the JPEGs inside more aggressively.

The harmless part: cleanup

Every PDF carries some baggage that is not content: revision history, unused objects, duplicated resources, over-detailed metadata. Stripping these is entirely lossless โ€” the document looks pixel-identical. On its own it typically saves 5-15%. A modest win, but with zero risk.

The irreversible part: image re-encoding

This is where the real savings come from, and where you make a trade-off. Two separate settings are at work, and they are commonly confused:

  • Resolution (DPI): how many pixels the image contains. Dropping a 600 DPI scan to 150 DPI shrinks the file dramatically. 150 DPI is plenty for on-screen reading; print needs at least 300 DPI.
  • JPEG quality level: how much information is kept at the same pixel count. 90% is usually indistinguishable; below 60% smudging becomes visible.

Why are scanned documents so large?

A scanner does not save the text on a page as text โ€” it saves it as a photograph. A 10-page scanned contract is really 10 high-resolution photos. That is why the same content produced digitally might be 200 KB while the scanned version is 40 MB. Scanned documents are where compression pays off most โ€” often over 90%.

Things to try before compressing

  • Delete pages you do not need โ€” blank pages, covers, repeated appendices
  • If you are merging several PDFs, compress each one before merging
  • If the document is scanned, set the scanner to 200-300 DPI from the start; 600 DPI is overkill for most documents
  • If colour is not needed, scan in greyscale โ€” it cuts size by roughly two thirds

All the PDF tools below run inside your browser. Because your document is never uploaded to any server, you can safely use them for sensitive files like contracts or identity documents.

Tools Mentioned in This Article