Compress PDF
Shrink a large PDF by re-rendering pages — works best on scanned documents.
How to Use
- Drop a PDF onto the drop zone, or click to browse.
- Read the note about text becoming non-selectable and decide whether that is acceptable for your file.
- Pick a compression level — stronger compression means a smaller file and softer image quality.
- Click Compress, compare the before and after size, then download if you are happy with the result.
When to Use This
- Shrinking a phone-camera scan of a signed form down to the few-megabyte ceiling that government and bank upload portals tend to enforce.
- Reducing a long scanned brochure for email, where the content is already images and nothing searchable is lost by re-encoding them.
- Compressing a scanned lease to attach to a ticketing system that rejects large files, in a document whose text was never searchable anyway.
- Checking whether compression is worth doing at all, since the before and after sizes are shown and the tool says plainly when the file grew.
Frequently Asked Questions
About This Tool
Most of a PDF's size sits in a handful of image streams. A scan is one large image per page; a text document is a few kilobytes of glyph positions plus the font programs. There is no general way to make the second kind much smaller, which is why honest PDF compression is specific: re-encode the images, and accept that whatever was text becomes pixels too.
This tool takes that trade in full. pdf.js renders each page to a canvas at 150, 110 or 80 DPI depending on the level you pick, the canvas is encoded as a JPEG at quality 0.85, 0.65 or 0.45, and pdf-lib places that single image on a new page matching the original's exact point size. Before and after sizes are compared and reported, including when the result came out larger.
Notes & Limits
- Every page becomes a JPEG, so selectable text, links, form fields and annotations are absent from the output; keep the original if any of those matter.
- At the strongest level pages are rendered at 80 DPI, well below print quality, and small type in a scan can become difficult to read.
- Pages are rendered one at a time in your tab, so several hundred pages is slow even at the lightest setting and can run the tab out of memory.
- There is no option to recompress only the embedded images and keep the text layer, which is why this tool has nothing useful to offer a text-based PDF.