The files people most need to compress — legal filings, financial statements, HR paperwork — are exactly the files that shouldn't be handed to a random web service. This compressor takes the server out of the equation: it parses your PDF locally, strips the metadata and dead objects, and repacks everything into the compressed object streams that PDF 1.5 introduced, all inside your browser. One honest note: it won't dramatically shrink image-heavy scans, because it never re-encodes your document's images — what it removes is structural bloat.
Drop your file here or click to browse
.pdf — up to 200 MB. Processed locally; never uploaded.
How to compress a PDF Without Uploading It
Drop the PDF onto the converter above.
The header is checked and the file size measured against the 200 MB cap.
Click Compress — parsing and re-serialization happen on your device.
Read the before/after sizes shown with the result.
Download if the savings are worthwhile; your original is untouched either way.
Frequently asked questions
Why do server-based compressors worry people?
Because compression requires the full document: upload is mandatory, and you're trusting the operator's deletion policy, their security, and their jurisdiction. In-browser compression removes the third party from the trust equation entirely.
What exactly gets removed during compression?
Document metadata (author, producer, timestamps), orphaned objects left by editing tools, and inefficiently packed structures — then everything is rewritten into compressed object streams. Pages, text, and images are not altered.
How much will my file shrink?
Metadata-bloated office exports often drop 10–20% or more; image-heavy scans may barely move, since their bulk is already-compressed photos. The tool reports the real result before you download, so there's no guessing.
Is stripping metadata good or bad?
For sharing, mostly good: producer strings and author names are a classic metadata leak. If you need the metadata preserved for records management, keep the original — compression here always writes a new file.