Compress PDF
betaMake a PDF smaller without changing what it says — pictures re-encoded, waste removed
How it works
Open the PDF and choose a preset: Light keeps every picture as it is and only tidies the file; Balanced re-encodes pictures to at most 150 pixels per inch at JPEG quality 80 %; Strong goes to 96 ppi and quality 60 % and removes metadata. Each picture's target size comes from how large it is drawn on the page, so a 4000-pixel photo shown two inches wide loses what no screen or printer would show. Pictures already small, and those in JPEG 2000, JBIG2 or fax encodings, are left alone. The status line reports the before and after sizes and what was done; undo brings the original back. Text, vector drawings, forms and comments are untouched.
Questions
Why did my file barely shrink?
It was probably mostly text, which is already compressed, with few or no pictures. Compression gains come from pictures; a Word or LaTeX export with none will shrink only by the tidying steps.
Does compressing lose quality?
Pictures are resampled and re-encoded as JPEG, so some detail goes — at 150 ppi nothing visible on screen, at 96 ppi a little in print. Text and drawings are vector and lose nothing.
Is my file uploaded?
No. Everything runs in your browser — pdf.js draws the document and pdf-lib rewrites it in a background thread — and Save writes to your disk or downloads a copy. Nothing leaves your device.