How this compressor works — read this first
This tool reduces PDF size by rendering each page to an image and rebuilding the document from those images. Every page is drawn at 1.5× screen resolution, encoded as a JPEG at the quality you choose, and placed on a fresh page of the same size. The result is often a fraction of the original size — but it is a different kind of PDF. Text is no longer selectable or searchable, links stop working, and zooming in far enough shows JPEG pixels rather than crisp vector text. That trade-off is ideal for documents that are already scans or photos, and for anything that only needs to be viewed or printed. It is the wrong tool for a document you will edit, search, or submit where text extraction matters.
Choosing a quality level
| Setting | JPEG quality | Typical size | Use when |
|---|---|---|---|
| High compression | 40% | Smallest — often 5–10× smaller | Email attachments, archiving, screen-only viewing |
| Medium compression | 60% | 3–6× smaller | General sharing; good balance for scans |
| Low compression | 80% | 2–4× smaller | Documents with photos or fine detail that will be printed |
Sizes are relative to a typical scanned document. A PDF that is already made of compressed JPEG scans may shrink less; a PDF exported from a design tool with uncompressed images or many embedded fonts may shrink dramatically.
What makes PDFs large
- Images — almost always the main cause. A 12-megapixel phone photo embedded at full resolution is 3–5 MB by itself, even though a page only needs about 1 MB of pixels to print well.
- Scans at high DPI — 600 dpi colour scans of text pages are ten times larger than 200 dpi greyscale scans and look no better on screen.
- Embedded fonts — full font files (rather than subsets) can add 100 KB to several MB each; CJK fonts are the largest.
- Leftover data — PDFs edited many times accumulate unused objects and old revisions that a plain re-save leaves in place.
- Transparency and vector complexity — maps, CAD drawings and charts with tens of thousands of paths can be large while containing no images at all.
When to use a different approach
- Need selectable text? Use the “Reduce file size” or “Optimize” export in Acrobat, macOS Preview, or LibreOffice, which downsample images while keeping text as text.
- Just a few oversized images? Compress the images before creating the PDF, or re-export from the source document with a lower image quality setting.
- Need to hit an exact size limit? Try medium first, check the result, and step down only if needed. Each level roughly halves the size.
- Vector-heavy drawings? Rasterizing them can make them larger, not smaller. Flatten or simplify in the original application instead.
Common uses
- Emailing scanned contracts or ID documents that exceed a 10–25 MB attachment limit.
- Uploading to portals with strict size caps (government forms, university applications, job sites).
- Shrinking phone-scanner output, which is typically far larger than it needs to be.
- Archiving old paper records where searchability was never available anyway.
Privacy
Rendering and re-encoding happen entirely in your browser using PDF.js and pdf-lib. The file is never uploaded, which is a meaningful difference from most online compressors when the document contains personal or financial information.