Compress a PDF without turning the page into a picture
Most compressors shrink a file by rasterising every page. This one makes a lossless pass first: redundant objects and duplicate fonts go, text and vectors stay exactly as they were.
Where the megabytes actually are
PDF bloat usually comes from three places, and only one of them is pictures. First, orphaned objects: an editor deletes a page but leaves its images and fonts in the file, because nobody garbage-collects them. Second, duplicate font subsets: each edit pass embeds another copy of the same 14 pt serif. Third, genuinely large images — a 600 DPI scan pasted at 40% size still carries all 600 DPI.
A lossless pass fixes the first two and cannot fix the third. That is not a defect; it is arithmetic. If your file is 80 MB of photographs, no structural cleanup will make it 5 MB without changing the pixels, and any tool claiming otherwise is re-encoding your images and calling it compression.
How compression works here
- Add the PDFThe preview reports input size and page count before you commit.
- Check the estimateYou see what will be rewritten and what will be left alone, so a small saving is explained rather than mysterious.
- Export and compareKeep both files. If the saving is not enough, the images are the reason — and that is a different job.
Why there is no “strong” button
A strong-compression switch is easy to ship and hard to use well: it silently downsamples every image, so a page that looked fine on screen prints soft, and a scanned signature stops being legible. We would rather have one honest level. If you need a smaller file than lossless can give, the right move is to reduce image resolution deliberately — export the pages to JPG at 150 DPI with PDF to JPG and rebuild, knowing exactly what you traded away.
| Item | Behaviour |
|---|---|
| Method | Structural rewrite, compressed object streams |
| Text and vectors | Untouched — still selectable, still sharp |
| Images | Not re-encoded, not resampled |
| Typical saving | 5–25% for edited or exported files; near 0% for image-heavy scans |
| Reversibility | Original file is never modified, so you can always go back |
FAQ
Is it really lossless?
The content is preserved exactly — text, vectors, image pixels and metadata. The bytes differ because the file is rewritten, so a checksum comparison will not match, but the rendered page is identical at any zoom.
Why did my file only shrink by 3%?
Because it was already well built, or because it is mostly photographs. Lossless compression removes waste; it cannot invent savings that are not there. Check the image count before assuming the tool failed.
Will compressed files still open in Acrobat and Preview?
Yes. The output is a standard PDF 1.7 file with object streams, which every current reader supports. Very old readers (pre-2010) may struggle with compressed streams.