Loading…
Compress PDF files to reduce file size while maintaining quality. Optimize PDFs for web sharing and storage.
Honest expectations: this tool re-serializes the PDF with object streams and optionally strips metadata. Gains are typically modest (often a few percent) and come entirely from removing structural overhead. It does NOT re-encode or downsample embedded images, which is where most PDF weight lives, so image-heavy files may barely shrink. Think of it as optimize and clean, not magic compression.
The PDF compressor re-saves a PDF in a more compact form to trim the structural overhead a document accumulates as it is created and edited. It loads the file with pdf-lib, rewrites it using object streams — which pack the internal indirect objects together more efficiently — and can optionally strip the document metadata at the same time. Add a file, choose whether to remove metadata, and download the optimized copy.
Everything happens inside your browser. The PDF is read locally, re-serialized in memory, and written back to your device, so nothing is uploaded to a server. That keeps sensitive files private and lets the tool work offline once the page has loaded.
Be realistic about the results. This is an optimize-and-clean pass, not aggressive compression: it does not re-encode or downsample the embedded images where most of a PDF's weight usually lives. Gains are typically a few percent and come from tidying structure and dropping metadata, so an image-heavy file may barely shrink and an already-optimized file may not shrink at all. For large reductions on scanned or photo-heavy PDFs, a tool that resamples images is the better choice.
Usually only a few percent. The tool removes structural overhead and optional metadata but does not re-encode embedded images, so image-heavy files may barely change and already-optimized files may not shrink at all.
No. The PDF is loaded, re-serialized, and saved entirely in your browser, so it never leaves your device and the tool works offline once loaded.
If a PDF is already well optimized there is little structural overhead left to trim. Most of the remaining size is likely in embedded images, which this tool does not resample.
No. Text and images are preserved as-is; only the internal structure is rewritten more compactly and, if you choose, the metadata is removed. Nothing is downsampled or re-encoded.