“Your file exceeds the size limit.” PDFs bloat fast — a few phone-camera scans and you’re over an email or upload cap. Here is why it happens and how to fix it.
We measured this ourselves. Our 334.1 KB test document came out at 167.6 KB — exactly half. But look at what paid for it: our tool renders every page to a JPEG and rebuilds the PDF from those images, so the text layer is gone. Four of the five test pages are plain text, and after compressing they are photographs of text — not selectable, not searchable, not copyable. That makes this right for scans and image-heavy files you only need to view or send, and wrong for anything someone will search. See all our measurements
Why PDFs get big
Almost always, it’s the images inside. A scanned or photographed document stores full-resolution pictures of each page. Text-only PDFs, by contrast, are already tiny. So:
- Scanned / image-heavy PDF → big, and very compressible.
- Text/vector PDF → already small; little to gain.
Our own test document makes the cause obvious. It is 334.1 KB in total, and splitting it shows why: page 3 alone — a full-page photograph — accounts for 331.2 KB of that, while a plain-text page is 1.4 KB. PDFs get big because of images, essentially never because of text.
Ways to shrink a PDF
- Recompress the images — re-render each page at a sensible resolution and JPEG quality. This is where the big savings come from for scans. The trade-off: pages become images, so text is no longer selectable.
- Remove unneeded pages — split out just the pages you need.
- Don’t over-scan — if you control the scanner, 150–200 DPI is plenty for documents.
Compress it in your browser — nothing uploaded
Contracts, IDs and financial statements are exactly the files you should not upload to a random site. This runs entirely on your device:
- Compress PDF — drag a quality slider and watch the size drop; best for scanned or image-heavy PDFs.
Also useful nearby:
- Merge PDF / Split PDF — combine or trim before compressing.
- Building a PDF from photos? See How to combine images into a PDF — and compress the images first for a smaller result.
What to expect
A scanned PDF can often drop to a fraction of its size at a quality that still prints and reads fine. A text PDF that’s already small won’t shrink much — and that’s normal.
Concretely, at the page default our 334.1 KB document came out 167.6 KB (-50%). But look at where that came from: the photo page was 331.2 KB of the original, so halving the file is really just re-encoding that one page. A text-only PDF has almost nothing to squeeze — which is why “compress” is not a universal fix.
Fonts and other hidden weight
Images are the usual culprit, but not the only one. Embedded fonts add weight too — every typeface is packed inside so the document renders identically on any machine. Files that mix several fonts, or embed full (non-subset) fonts, carry kilobytes you never notice until they add up. Other quiet contributors: embedded ICC colour profiles, high-resolution vector artwork, and leftover revisions from repeated “Save As” chains. None of these squeeze the way a photo does, which is why a text-heavy report sometimes refuses to shrink.
How image resampling works
When a tool “recompresses images” it does two separate things: downsampling and re-encoding. Downsampling throws away surplus pixels — a page scanned at 600 DPI but only ever read on screen has roughly four times more pixels than it needs. Re-encoding then stores those pixels as JPEG at a chosen quality. The two multiply: halving the resolution cuts the pixel count to a quarter before JPEG even begins, which is why the biggest savings come from resolution, not quality alone. If you’re building the PDF from photos, reduce the image file size first.
Picking a level on our tool
Our Compress PDF page gives you one control: a quality slider from 30 to 90, defaulting to 60. There is no DPI setting, because the render resolution is fixed at 1.5× the page’s own point grid (roughly 108 DPI) — enough to read comfortably on screen, not enough for archival printing.
At that default, our 334.1 KB five-page test document came out at 167.6 KB. Pushing the slider higher keeps more detail and saves less; lower does the reverse. What does not change at any setting is the trade-off: every page becomes a JPEG image, so pick a level for how the result will be looked at, and use a different approach entirely if the text has to stay searchable.
Common mistakes
- Expecting miracles from a text PDF. If it’s already vector text there’s little to remove — a sign it was never the problem.
- Re-compressing an already-compressed file. Each JPEG pass adds artefacts; compress once, from the best original.
- Silently losing the text layer. Image recompression flattens pages, so search and copy stop working. Keep the original if you need selectable text.
FAQ
Does compressing a PDF reduce quality?
For scans, a little — you trade some sharpness for a far smaller file, usually invisibly at sensible settings. Text and vector PDFs are untouched because nothing is rasterised.
Why is my PDF still large after compressing?
It’s probably already efficient: mostly text and vectors, with little image data left to squeeze. Splitting out unneeded pages is then the better lever.
Is it safe to compress private documents online?
Only if the tool runs in your browser. A browser-based PDF compressor never uploads the file, so contracts and IDs stay on your device.