Compressing a PDF restructures how the document is stored internally, removing redundant structural overhead. It does not re-encode or downscale embedded images. A text-heavy PDF with a messy internal structure can shrink noticeably; a PDF that's already mostly large photos will shrink much less, because there's less structural redundancy to remove.
What "compression" actually optimizes
A PDF isn't just visible pages — internally it's a structured file of objects (fonts, page descriptions, images, metadata) linked together by a cross-reference table. Older or repeatedly-edited PDFs often carry redundant structure: duplicate objects, an inefficiently laid-out cross-reference table, or leftover data from prior edits.
Restructuring that internal layout — grouping objects more efficiently — can meaningfully shrink a file without touching a single pixel of its visible content. That's the mechanism at work: structural cleanup, not visual re-encoding.
Why image-heavy PDFs shrink less
If most of a PDF's size comes from large embedded photos or scanned pages, structural cleanup has a much smaller effect — the images themselves are untouched, and they're usually the majority of the file's weight. A PDF built from scanned pages will typically shrink far less than a PDF built from formatted text and vector graphics.
If your goal is specifically to shrink an image-heavy PDF, compressing the source images before placing them in the document — or via an image compressor — addresses the actual weight, rather than the PDF's structure.
Don't expect a fixed percentage. A poorly-structured, text-heavy PDF might shrink substantially; an already-efficient or image-dominated PDF might barely change. Both are normal outcomes of the same process, not a sign something went wrong.
Compressing privately, in the browser
Compression here happens by loading the PDF into a JavaScript PDF library running in your browser tab, then rewriting its internal structure and saving the result — no upload, no server round trip. You can verify this directly: open your browser's Network tab before compressing, and no PDF data appears in any outgoing request.
Limitations
A password-protected PDF can't be compressed directly — its internal structure isn't readable without the password. Remove the password first. And because compression here doesn't touch embedded images, a scanned-document PDF should be treated separately if file size for that specific case is the goal.
Check what's making the file large
Text-heavy structure responds well to compression; large embedded images mostly don't.
Unlock protected files first
Encrypted PDFs need their password removed before the structure can be rewritten.
Compress, then compare
Check the before/after size shown, and open the result to confirm nothing visible changed.
Frequently asked questions
Does PDF compression re-encode the images inside the file?
Not with structural compression. Restructuring a PDF's internal object streams shrinks redundant document structure, not the embedded images themselves — a scan-heavy PDF full of large photos will shrink less than a text-heavy one.
Why did my PDF barely get smaller?
If the file is already efficiently structured, or its size mostly comes from large embedded images, structural compression has little redundant overhead left to remove. That's expected, not a failure of the tool.
Is compressing a PDF in the browser private?
Yes — the file is read and rewritten locally using a JavaScript PDF library, and never uploaded. You can confirm this in your browser's Network tab.
Can I compress a password-protected PDF?
No — an encrypted PDF's internal structure can't be read and rewritten without the password. Remove the password first, then compress.
Primary reference
- ISO 32000-1 (PDF 1.7 specification), §7.5.7 — cross-reference and object streams, the structural mechanism PDF compression optimizes.