Merge, compress, and shrink PDFs right here in your browser. Free, instant, and completely private — nothing is ever uploaded.
Repacks the PDF's internal structure without touching pages, text, or image quality. Great for text-heavy documents; photo-heavy PDFs shrink less — use Shrink to size for those.
Only one file at a time, because the tool reports a precise before-and-after comparison for it. Merge first if you have several.
The document is parsed, rebuilt, and written back out. On a typical report this takes well under a second; a 500-page manual takes a few.
You get the original size, the new size, and the percentage saved. If the saving is negligible, keep your original — nothing is lost either way.
A PDF is not one continuous stream of data. It is a container holding thousands of numbered objects — fonts, images, page descriptions, metadata — plus a cross-reference table that says where each object lives. Producers such as scanners, Word, and export plugins are often careless with that container. Lossless compression cleans it up:
What it deliberately does not do: re-encode images, downsample resolution, subset or drop fonts, or rasterise anything. Every glyph, every pixel, and every vector path in the output is bit-for-bit what it was in the input. Text stays selectable, links keep working, and print output is identical.
Anyone promising a fixed percentage is guessing. The realistic ranges below come from where the bytes are in each kind of document.
| Type of PDF | Typical saving | Why |
|---|---|---|
| Word or LibreOffice export | 5–25% | Loose object structure and duplicated font data |
| Report edited and re-saved many times | 15–40% | Stale incremental updates and orphaned objects |
| Slide deck exported to PDF | 5–20% | Repeated template graphics on every slide |
| Scanned document | 0–5% | Nearly all bytes are already-compressed images |
| Photo portfolio or brochure | 0–3% | Same reason — JPEG data cannot be deflated further |
| PDF from a modern optimising producer | 0–2% | Already packed properly; nothing left to reclaim |
A result of zero is a legitimate outcome. It means the file was well built.
If you need a specific number — under 10 MB for a portal, under 25 MB for an email — lossless compression cannot promise it, because it has no lever to pull once the structure is clean. At that point you are choosing between three options:
| Your situation | Use |
|---|---|
| The file is slightly too big and quality is non-negotiable | Compress (lossless) |
| A contract, invoice, or anything that may be printed or searched | Compress (lossless) |
| You must be under an exact megabyte limit today | Shrink to size |
| Scanned pages that only need to be readable on screen | Shrink to size |
| You are not sure | Try compress first — it costs nothing and never degrades the file |
Set an exact target in megabytes and let the shrink tool find the best quality that fits underneath it.
Yes. Compression runs in JavaScript inside your browser tab. The PDF is read into memory on your device, rebuilt there, and never uploaded, stored, or transmitted. Close the tab and it is gone.
Yes. Lossless compression never rasterises anything. Text, fonts, links, and bookmarks come through untouched, which is exactly what separates it from the shrink tool.
Because it was already efficiently packed, or because almost all of its bytes are image data that is already compressed. A file can occasionally grow by a fraction of a percent when the rebuilt structure adds a little overhead to an already-optimal layout. Keep the original in that case.
No. Document metadata is preserved. If you need to strip author names, timestamps, or producer strings, do that in a PDF editor before or after compressing.
Usually not. Encrypted documents cannot be parsed without the password and will fail with an error. Open the file in a reader, remove the protection, save a copy, and compress that.
No fixed limit — it depends on your device's available memory. Very large files may take a few seconds and briefly use a lot of RAM, since the whole document is held in memory while it is rebuilt.
Compress is lossless: it repacks the file's internal structure and changes nothing you can see. Shrink is lossy: it re-renders pages as compressed images to reach a target size, saving far more space at the cost of selectable text and fine detail.