Compress PDF: Some Files Come Out Bigger
This tool shrinks a PDF by re-baking each page as an image. That works extremely well on scans and photo-heavy documents and works backwards on text. Knowing which kind of file you have is the whole decision.
A PDF is usually large for one reason: the images inside it are stored at a higher resolution than anything needs. Scanned documents, slide decks full of photographs and anything assembled from phone camera shots all fit that description.
This tool draws each page, saves that rendering as a JPEG, and assembles a new PDF from those images. Whatever the original contained, the result is a stack of pictures.
Your text becomes an image
This is the consequence that matters. Afterwards the text cannot be selected, copied or searched. It still looks like text, but it is a photograph of text.
For a contract, a paper or a CV β anything you will later need to search inside β this is the wrong approach. It also puts the content out of reach of anyone using a screen reader.
For a scan, on the other hand, nothing is lost. A scanned PDF is already images and was never searchable to begin with.
Why text-heavy documents grow
Text is stored remarkably efficiently inside a PDF. It amounts to instructions saying which glyph of which font goes where, so a full A4 page of text often costs only a few kilobytes.
Rendering that page as an image changes the economics completely. To stay readable it needs to be at least 900 pixels wide, and a JPEG that size runs to hundreds of kilobytes. A page that cost a few kilobytes now costs dozens of times more.
So when the result comes out larger than the original, this tool says so and tells you to keep your original. It does not pretend to have saved you anything. The problem is not that nothing can be done β it is that this technique cannot do it.
What the quality settings control
The scale decides how large each page is rendered; the quality decides how hard the JPEG is compressed.
| Setting | Render scale | JPEG quality | Suits |
|---|---|---|---|
| Maximum compression | 1Γ | 40% | Scans that will only be viewed on screen |
| Balanced | 1.5Γ | 65% | Most files |
| Minimum compression | 2Γ | 85% | Anything being printed, or with small type |
Files this suits
- β’Documents scanned with a phone β almost always at more resolution than needed.
- β’Slide decks carrying several photographs.
- β’Image-heavy files hitting an email attachment limit.
- β’PDFs that are already images, where there is no searchable text to lose.
Files to keep away from it
- β’Text documents β contracts, reports, CVs. They grow and stop being searchable.
- β’Anything you will need to copy from or search later.
- β’Print-ready artwork, where rasterising costs you sharpness.
- β’PDFs with fillable form fields, which do not survive the process.
Extended FAQ
My file got bigger.
It was almost certainly text-heavy. Pages that cost a few kilobytes as text now cost hundreds as images. This technique cannot shrink that file β keep the original.
I can no longer select the text.
That is expected: the pages are images now. If you need search and copy, this method is not usable for that file.
Does it handle CJK documents safely?
Fonts become pictures, so there are no font-embedding or tofu problems. Hangul and Han characters have many strokes and blur at low resolution, though, so use Balanced or better rather than Maximum compression.
Is my file uploaded?
No. Everything happens in your browser, which is also why large files take time and depend on your device.
Can I process several files at once?
Yes. Process a batch and you can download them together as a zip.