PDF Compressor

What this does

Rewrites a PDF in a tighter file structure and downloads it as name-compressed.pdf, with a line underneath showing the old size, the new size and the percentage change. It is a lossless pass: text, fonts, vector artwork and embedded images come out exactly as they went in. Only the packaging changes. The file is parsed and rewritten by pdf-lib inside your browser tab and is never uploaded, so a signed contract or a bank statement is as private here as on your own disk.

Two things shrink the file. The document is saved with object streams, which pack the thousands of small dictionaries a PDF is made of into compressed blocks instead of writing each one as plain text. And with Strip metadata ticked, the default, the title, author, subject, keywords, creator and producer fields in the info dictionary are blanked.

How do I compress a PDF?

1. Choose the PDF with the file picker.
2. Leave Strip metadata ticked unless you need the title and author fields to survive.
3. Press Compress & download. The rewritten file downloads with -compressed added to its name.

How much smaller will the PDF be?

It depends on how the file was written. PDFs from older printer drivers, scanners and export filters that never used object streams commonly drop by 10–30%. Files saved many times with incremental updates (some editors append a new revision on every save) shed all the dead earlier revisions, which can be dramatic. A PDF from a modern tool that already uses object streams may lose only a few kilobytes, and can occasionally come out marginally larger; the result line shows a plus sign when that happens, and you can keep the original.

The one thing this tool cannot shrink is image data. A scanned document is mostly JPEG bytes that are already compressed, and repacking the structure around them barely moves the total.

Why doesn't it recompress the images?

Because doing that well means decoding every image, downsampling it, encoding it again at lower quality and splicing it back into each page's resource tree. That is lossy, full of judgement calls (a photo can take heavy downsampling, a scanned signature cannot), and far heavier than a browser tab should carry. If your problem is a 40 MB scan, fix the images before they become a PDF: run them through the image compressor, then rebuild the document with image to PDF. For an existing image-heavy PDF, a desktop tool such as Ghostscript is the right instrument.

Does compressing change how the PDF looks or works?

No. Text stays selectable and searchable, fonts are untouched, page size, order and rotation are preserved, and nothing is rasterised. Because the whole document is loaded and re-saved rather than copied page by page, document-level structure survives too: outline bookmarks, links, annotations and form fields all stay in place. The metadata option only clears descriptive fields and leaves any XMP metadata stream alone, so treat it as tidying rather than full anonymisation.

Is there a file size limit?

No fixed cap. The original bytes and the rewritten copy are both held in memory during the save, so the ceiling is your device's RAM. If a very large scanned PDF fails on a phone, try a desktop browser, or split it with the PDF splitter and compress the parts.