How to Compress PDF Files — and Why They Get So Big

Compress a PDF in three steps

1. Open Compress PDF — processing happens locally in your browser.
2. Drop your file and choose a compression level. Start with medium: it downsamples images but keeps text sharp.
3. Check the result, especially the text pages. If the size is still over your limit, go one level stronger and compress again.

What actually makes a PDF heavy

Text is almost never the problem: 50 pages of plain text weigh well under 1 MB. The weight lives in images — and most heavy PDFs are heavy for exactly one reason:

• Scanned documents. A scanner or phone scanning app stores each page as a full-resolution photo, often 300 DPI in uncompressed or lightly compressed form. A 20-page scan at 4 MB per page is an 80 MB PDF before you blink.
• Embedded photos. Brochures, portfolios and reports with exported-at-maximum-quality images.
• Duplicated resources. The same logo or font embedded once per page by clumsy export tools.
• Fonts and extras. Full embedded fonts can add megabytes, though subsetting usually keeps this small.

This explains the golden rule of PDF compression: if your PDF is image-heavy, image compression is where the savings are. Text-only PDFs are already small and barely benefit.

Choosing a compression level

• Light: barely touches images; use it when a PDF is just over a limit and quality is critical (print-ready material).
• Medium: downsamples images to screen-friendly resolution (~150 DPI). The right default for scans and documents destined for screens — text stays vector-sharp because text isn't an image.
• Strong: aggressive downsampling for hard caps (email limits, strict portals). Expect visibly softer images, but a 45 MB scan becoming a 4 MB PDF is normal at this level.

One subtlety worth knowing: text in a PDF created from a word processor is stored as vector text and stays perfectly sharp at any compression level. It's scanned pages — which are photographs of text — that soften. If your PDF must be text-searchable and text-crisp, compress images but avoid tools that rasterise whole pages.

Realistic size expectations

Scans routinely shrink by 90% because camera-resolution pages carry enormous redundancy. Text-only documents may only lose 10–30% — that's not the tool failing, it's the file already being efficient.

What to check after compressing

A compression pass is only done when you've verified the result — thirty seconds that save re-sending documents:

• Text still selects and searches. Compression preserves the text layer; if Ctrl+F suddenly fails, the tool rasterised pages and you should pick another one (or another level).
• Pages render in the right order and none are blank — aggressive levels occasionally drop near-empty pages that mattered (a signature page, a blank form divider).
• Images are acceptable at display size. Judge screenshots at 100% zoom on the smallest screen your audience uses; softness that vanishes at laptop size may still look rough on a phone.
• File properties updated. Confirm the new size in the file manager, not just in the tool's estimate — and keep the original untouched for future re-compression.

Finally, resist the urge to compress a compressed file repeatedly: each pass re-encodes the embedded images and compounds softening. Compress once from the original at the right level; if the level was wrong, go back to the source.

Alternatives when compression isn't enough

If you're still over a hard limit: split the PDF and send parts (the Split PDF tool), re-scan smarter (200 DPI greyscale instead of 300 DPI colour produces far lighter source files), or convert the content to plain text with PDF to Text when the recipient only needs the words. And if you regularly email documents, compressing once on your side beats asking every recipient to deal with a 40 MB attachment.