Compress PDF

Where a PDF's size comes from

In most large PDFs almost all the bytes are pictures: scanned pages, phone photos, screenshots. Text and vector drawings take very little space. So this tool goes after the pictures: each one is scaled down to a sensible resolution and saved again as a JPEG at a lower quality, while text, fonts and drawings are left exactly as they were.

Pick a level:

• Light keeps photos up to 2400 pixels on their long side at high quality, for print.
• Recommended (1800 px) suits screen reading and most upload forms.
• Strong (1200 px, lower quality) gives the smallest file, still readable on screen.

A picture is only replaced when the new version is clearly smaller, and if the whole file doesn't get smaller you get the original back. You can compress up to five PDFs at a time.

Nothing is uploaded

Your PDF is processed by your own browser, so confidential documents never reach a server and there is no daily limit.

The PDF is opened in a Web Worker with pdf-lib. Every JPEG image and every uncompressed-colour RGB or greyscale image is decoded by the browser, resized on a canvas and encoded as JPEG at the chosen quality (0.80, 0.65 or 0.50). The smaller version replaces the original inside the PDF; images with transparency masks, CMYK colour or special encodings are kept as they are. The document is then saved with compressed object streams.