Reduce PDF file size without recompressing images
This tool makes a PDF smaller by rebuilding it, not by shrinking its pictures. It removes objects nothing points to any more and, if you allow it, optional metadata links. The result is lossless, so a PDF made of scans or large photos may shrink very little or not at all, and the page shows the exact before and after sizes so you can see.
Runs in your browser
Free, no sign-up. The PDF is read and rebuilt in your browser; its bytes are never uploaded. The tool records only anonymous actions such as 'compress' or 'download', never file names or sizes.
PDF Compressor
Compress PDF files in your browser with a safe lossless rebuild.
Files
0
Saved
0%
Output
0 B
This optimizer validates PDF structure, removes safe metadata references, drops unreachable incremental-update objects, and rebuilds the cross-reference table locally. Image-heavy PDFs that are already optimized may not shrink much without a full PDF renderer.
PDF queue
0 completed of 0 files
Upload one or more PDF files to see compression results here.
How to use the PDF Compressor
- Drop PDFs onto the upload area or click it to choose files: up to 10 files of 100 MB each.
- Under Compression settings, keep or untick Remove metadata references, and keep or untick Keep original if rebuild is larger.
- Press Compress All (or compress a single file in the PDF queue).
- Read the Files, Saved and Output figures, and the message on each row: a file that could not shrink says so.
- Download each optimised PDF, then open it to check it looks right before you discard the original.
Worked example
Sample data
A report PDF that was edited and saved ten times in an editor that appends changes, with old revisions still stored inside the file.
Rebuilt file: the unreachable objects left by earlier saves are dropped and the index is rewritten. How much that saves depends on the file; the page prints the real figure for each PDF.
A scan exported once from a phone is the opposite case: it is mostly image data, already stored once, so a lossless rebuild has almost nothing to remove.
Good to know
If you need a PDF under a fixed size, such as 1 MB, and the file is mostly scanned pages, this tool is the wrong one: it never downsamples embedded images. Re-export the document at a lower image resolution from the app that created it, or compress the page images separately first.
How it works
What this compressor actually does
The tool reads the PDF's internal objects, removes objects that nothing points to any more (typical after repeated 'save' operations that append changes), strips optional metadata links when you allow it, and writes a clean cross-reference table.
It does not recompress images, subset fonts or convert colour spaces. That is why the result is safe and lossless, and also why a PDF made mostly of scanned pages or large photos may shrink by only a little or not at all.
Which PDFs shrink most
Files that were edited and saved many times in an editor that appends changes, or that carry leftover metadata, tend to benefit. Files exported once from a modern tool, and image-heavy PDFs, are usually already about as small as a lossless rebuild can make them.
If you need a big reduction on a scanned document, you need a tool that downsamples the embedded images; this one deliberately does not.
Metadata removal and the safety check
Remove metadata references drops the XMP /Metadata and /PieceInfo links from object dictionaries where doing so does not change the visible pages. It is not a redaction tool: document text, visible content and the classic info dictionary are not scrubbed.
With Keep original if rebuild is larger ticked (default), a rebuild that is not smaller is discarded and you get the original bytes back, so a download never becomes bigger.
saved % = (original size - output size) / original size x 100
Inputs, outputs and limits
Inputs
- PDF files (checked for the %PDF- header and an %%EOF marker)
- Setting: Remove metadata references (default on)
- Setting: Keep original if rebuild is larger (default on)
Outputs
- One rebuilt PDF per input, or the original when the rebuild is not smaller
- Original size, output size and percentage saved per file, plus totals
Limits
- Up to 10 files per batch and 100 MB per file.
- Encrypted or password-protected PDFs are refused.
- Lossless rebuild only: images, fonts and colours are not recompressed, so scanned or image-heavy PDFs may shrink very little.
- Not a redaction or sanitising tool.
Processing
- The PDF is read and rebuilt in your browser; its bytes are never uploaded. The tool records only anonymous actions such as 'compress' or 'download', never file names or sizes.
Frequently asked questions
Will my PDF look different afterwards?
No. The rebuild is lossless: pages, text, fonts and images are carried over unchanged. Only unreachable objects and, if ticked, metadata links are removed. Open the result and check it before discarding the original.
Why did my PDF not get smaller?
Most of its bytes are probably image or font data, which this tool does not touch. The row says so when the rebuild could not shrink the file, and with Keep original if rebuild is larger ticked you get your original bytes back.
Does it reduce scanned PDFs?
Usually only a little. A scan is mostly one compressed image per page. Shrinking it means lowering image resolution or quality, which this tool deliberately does not do.
Is the PDF uploaded anywhere?
No. The file is read and rebuilt in your browser tab and the result is offered as a download from memory.
Need this workflow inside your product?
We build calculators, converters and processing steps like this one directly into your web or mobile product, with your rules and branding.