Compress PDF reduces file size by targeting the actual source of most PDF bloat: embedded photos and scanned images. Rather than the common shortcut of rendering every page to a flattened image (which shrinks file size at the cost of your document's selectable text and sharp vector lines), this tool surgically locates each embedded JPEG image inside the PDF's own internal structure and recompresses just that image data, leaving every page's text and vector content completely untouched.

As soon as you drop in a file, a quick analysis pass reports how many embedded images it found and their combined size — genuinely useful information for deciding whether compression is worth running at all, not just a decorative loading step. If a PDF has no embedded images (a typed report or a purely vector diagram, for instance), that's shown plainly too, since there's nothing meaningful to shrink in that case and expecting a large size reduction would be dishonest.

Recompression itself happens across several background workers running in parallel, one per embedded image, so a PDF with dozens of photos compresses far faster than processing them one at a time on a single thread. Each image is decoded, redrawn at a size and JPEG quality appropriate to your chosen tier, and spliced back into the exact page position it came from — the surrounding page content stream (text positions, fonts, vector paths) is never rewritten.

We deliberately avoided two common shortcuts here: rasterizing whole pages (which destroys the text layer, as above) and shelling out to a general-purpose tool like Ghostscript compiled to WebAssembly, which would have meant shipping a large binary under a license that doesn't fit a lightweight, permissively-licensed tool site. The result is a compressor that's honest about what it can and can't shrink, and never damages the parts of your PDF it doesn't touch.