Find out where the size goes, then shrink the pictures or tidy the file. Vector text barely shrinks, and the tool says so before you start.
Your files never leave your browser — all processing is local.
Drop a PDF here
or click to choose a file. Read and rewritten in this tab, nothing uploaded
Vector text and exports from Word or slide tools barely shrink, because there are few pictures to re-encode. Re-encoding is lossy: check small print in the preview of your PDF reader. CMYK, JBIG2, JPEG 2000 and CCITT pictures are kept as they are. Digital signatures stop validating, and the page-picture mode is not offered here.
A phone scan of a three-page form can weigh 18 MB when the portal allows 5. Most of that weight is pictures, so this page reads the file first and lists every picture: its size, its effective resolution (pixels divided by the size it is printed at), its encoding and its share of the bytes. Only then does it estimate what re-encoding would save, and you decide whether to run it.
Level 1 rewrites the file with compressed object streams and drops objects nothing refers to. Pictures stay byte for byte the same, and the saving is usually small. Level 2, the default, re-encodes oversized JPEG and 8-bit pictures as JPEG at quality 75, scaled down to 150 or 200 DPI for where they are printed. A picture that would not get smaller is kept. Pictures with transparency masks, CMYK colour, other bit depths, or JBIG2, JPEG 2000 and CCITT encoding stay as they are, and the page gives the reason for each one.
Vector text and PDFs exported from Word or slide tools carry few pictures, so they barely shrink, and the analysis says so. Re-encoding is lossy: at 150 DPI small print can soften, so compare it in your PDF reader before you send it. Digital signatures stop validating after any change. There is no target size; the tool does what the picture rules allow and reports what it got.
No. The file is analysed and rewritten by JavaScript in this tab. Nothing is sent to a server, so a confidential form or scan stays on your device.
Most such files are mostly text and vector drawings with few or no pictures. The analysis shows the picture share before you run anything; when it is small, no setting can save much. PDFs exported straight from Word or a slide tool are the usual case.
Level 2 at 150 DPI for scans and photos you are going to send online. Level 1 when you do not want any picture changed at all. Either way, compare the result in your reader before you send it.
No. The tool does not promise a size. It shows an estimate before you run it and the real before-and-after sizes afterwards, and you can change the resolution and try again.
No. The compressed copy is a separate download, and the original stays where it was. Nothing is kept after you close the tab.
Join several PDFs into one file.
Cut a PDF into parts, or keep only the pages you need.
Turn photos and scans into a PDF, one picture per page.
Check and remove the author, producer and dates in a PDF.
Add a searchable text layer to a scanned PDF.
Shrink JPEG, PNG and WebP images before you put them in a PDF.