Hacks VitaeAll tools
Hacks Vitae Tools
Precision · Instant · Private

Compress a PDF, on your own machine

This rewrites how a PDF is put together, then saves a new copy. Nothing is re-encoded, and the file never leaves your device. It helps some PDFs a great deal and barely touches others — the section below explains which is which, so you can tell before you try.

SYSTEM ● ONLINE · LOCAL COMPUTE · ZERO UPLOAD
UNIT // PDF.COMPRESSLIVE
Drop a PDF file
Drag & drop or click to browse
—
Original
—
Compressed
—
Saved
What it does

It compresses the plumbing, not the pictures

// In short

A PDF is a pile of numbered objects: page descriptions, fonts, image data, and the table that says where each one sits. This tool reads yours, packs those objects together, and writes the file out again. Your pages, text and images come through unchanged, because none of them are re-encoded.

Why one PDF loses half its size and the next loses nothing

The saving comes from a feature the PDF standard calls an object stream. ISO 32000-1:2008 describes it as “a stream object in which a sequence of indirect objects may be stored”, and says its purpose “is to allow indirect objects other than streams to be stored more compactly by using the facilities provided by stream compression filters” (7.5.7). A companion table of contents for the file, the cross-reference stream, can be packed the same way (7.5.8).

Read the words “other than streams” carefully, because they decide the whole result. The same section lists what may not go into an object stream, and the first item is stream objects. Pictures are stored that way: the standard says an image XObject “is a stream object whose dictionary specifies attributes of the image and whose data contains the image samples” (8.9.1, NOTE 2). Page contents and embedded fonts are streams too.

So the method packs the file's structure and leaves every stream, pictures included, as it was. A document built up over many rounds of edits carries a lot of structure and little else, and can lose a noticeable share of its size. A scan is almost entirely image data, so there is very little left for this to pack — and that is the real reason it hardly moves, not a fault in the file.

FAQ

Frequently asked questions

Almost certainly because it is mostly images. This method packs the file's structure, and the PDF standard keeps streams, which is where image data lives, out of object streams (ISO 32000-1, 7.5.7 and 8.9.1), so a scan has very little for it to work on.
No. Nothing is re-encoded. The pages, text and images in the new file are the ones from the original.
Documents with a lot of structure and few photographs — forms, reports and contracts that have been edited and saved many times over.
No. It is read in the page and rewritten there. Nothing is sent, and nothing is kept after you close the tab.
That needs the images re-encoded rather than the structure packed, and it costs visible detail. Exporting at a lower resolution from whatever made the scan is usually the better route.
Mechanism

What gets rewritten, and what is left alone

The file is parsed, then written out again as a single current revision with compressed object and cross-reference streams. Everything the pages point at is carried across untouched.

Part of your PDFWhat happens to it
Cross-reference table and object streamsRewritten and compressed. This is where the saving comes from
Superseded data from earlier savesNot carried into the new revision
Photos and scanned pagesUntouched. Not re-encoded, not reduced in resolution
Embedded fontsUntouched. Not subset, even if the document uses six letters of one
Annotations, comments, attachmentsCarried over
Author, title, creation dateLeft exactly as found
Fast web view (linearisation)Not rebuilt, so a very large file may start displaying later online

Because metadata is deliberately left alone, this is not a redaction step. The author name and the software that made the file travel with it.

The case it cannot win

Scanned documents behave completely differently

Open a scan and look at one page. You cannot select the words, because there are none. Each page is a single large image, often already compressed by the scanner. There is almost no structure to squeeze, so a structural compressor has nothing to work with.

The size lives in the pixels, so that is where you have to go. The route that works, using tools on this site:

  1. If the scan is still a set of image files, shrink them first with compress image, or cut their width with resize image. Width is the bigger lever of the two.
  2. Rebuild the document with images to PDF.
  3. Run the result back through this page to tidy the structure.

If the scan only exists as a PDF, this tool cannot re-scan it for you. Reducing the scanner resolution and scanning again is usually faster than fighting the file.

Limits

When not to use this

  • The PDF is digitally signed. A signature covers an exact byte range of the file (ISO 32000-1, 12.8.1, ByteRange), and rewriting the file moves those bytes, so expect the signature to stop validating. Send the signed file as it is.
  • The PDF has a password. Encrypted files are not opened here. You get a short message saying the file could not be read. Remove the password in the application that set it, then come back.
  • You need PDF/A for an archive or a court filing. A rewritten file may no longer validate as PDF/A. Check before you rely on it.
  • You wanted metadata gone. It is preserved here on purpose. That is a different job, not a smaller file.
  • The file is damaged. Parsing is strict. A malformed PDF that a desktop reader quietly repairs on open may simply fail here. Open it in a reader and export a fresh copy first.
Worked example

An email bounced your 12 MB attachment

First, a detail that catches people out: mail attachments are usually sent base64-encoded, and base64 turns every three bytes into four characters (RFC 4648, 4), so the attachment grows by about a third on the way. A file sitting comfortably under the stated cap can still be refused.

  1. Drop the PDF here and read the Saved figure. If it is a decent double-digit percentage, you are done.
  2. If it is 0% or close, you have a scan or a photo-heavy file. Do not keep compressing it.
  3. Ask what the recipient actually needs. Two signed pages out of forty is a split, not a compression problem, and it will get you past any limit.
  4. If they need all of it, send a link from a storage service instead of an attachment.

For images rather than documents, start at the compression tools, which cover the formats where large savings are genuinely available.

Privacy

Why it matters that this one stays on your machine

Think about which PDFs you actually try to shrink. A signed contract. A payslip. A passport scan for a visa form. A letter from a hospital. A tax return. These are the documents that would cost you the most if they sat on a stranger's disk.

A conventional online compressor needs your file on its server to work on it. You are then trusting a retention policy you have not read, on infrastructure you cannot inspect. Here the library runs in the tab, the work happens on your processor, and the finished file is handed to your downloads folder. Nothing is stored, which also means there is no link to share and no copy to come back for. Reload the page and the result is gone.

The same applies to the neighbouring tools: merge, split and images to PDF all work the same way. So does everything else in the full list.

Sources

Where these facts come from

Last reviewed 21 October 2026

4 more

More documents tools

All 7 →

Everything here runs in your browser. Browse all the tools, or start from the home page.

Ask an assistant about this page

Opens in a new tab with this page and the question already filled in.