PDF will not compress: why, and what works instead

By the getPDF team · Published 11 October 2026

The short answer

A PDF will not compress when its weight is not in pictures: text, fonts and vector drawings are already compact, and a compressor leaves them as they are. It also will not shrink much if the pictures were compressed before, or if you picked a setting at or above their current resolution. Open the file in the inspector to see which case you have; then remove the pages you do not need, split the file, or grayscale a colour scan.

Find out what is inside, first

Retrying the same file in 3 more compressors usually gets the same answer 3 more times. Looking inside it takes seconds. Drop the file here:

Try it here, nothing is uploaded

Drop a PDF to inspect

PDF · any size

Read the section called Why it is this size. It lists the pictures, how many bytes they take, their share of the file, the picture formats, the largest picture, and “Everything else”. That share decides what can shrink.

Read the inspector's 'Why it is this size' box on the left; the fix that works is on the right.

Text-only files: the size is honest

A PDF typed in a word processor stores characters and a font, not pictures of pages. That is already compact. Our 500-page text test file is 117 KB, well under 1 KB a page, and it stays 117 KB on every setting; the compressor returns it unchanged with the line “the file has no pictures; its size is text, fonts and drawings, which this tool keeps as they are”. A 3-page text file goes from 2 KB to 1 KB, from tidying alone.

If a text-only file is still too big for the place it has to go, it is long, not bloated. Stop here and go to the fixes further down.

One exception: fonts. A file can carry large fonts, sometimes several of them in full. Our compressor does not subset or remove fonts, and its own page says so. The inspector lists every font, so you can see whether that is your case.

Already-optimised files: a second pass gains nothing

If the pictures were compressed before, by an earlier pass, a scanner app or a careful export, there is little left to take. Our compressor re-encodes a picture only when the result is at least 10 percent smaller, so an already compact JPEG is left alone, and a file where nothing shrinks comes back unchanged. We tried it: a 20-photo PDF went from 60.7 MB to 2.2 MB on the Email preset, and the 2.2 MB result, run through Email again, came back as it was, with the line “already as small as this tool can make it”.

How to recognise one: the inspector’s largest picture is modest in pixels (a portrait A4 page about 900 pixels wide is already at the Email preset’s 110 dpi), and the formats are mostly DCTDecode, which is JPEG.

Pushing harder with a size target still works, down to 50 dpi, but each pass on top of an earlier one stacks JPEG damage. If quality matters, go back to the original file and compress that once.

Three more cases that look like “will not compress”:

It was already under your size target. A file that already fits comes back untouched, at full quality, with a line such as “Already under 100 KB (34 KB): returned unchanged”. There was nothing to gain by re-encoding it.
You picked Print. Print keeps pictures at 300 dpi. On our 10-page test scan, made at 300 dpi, Print gives back the same 17.6 MB file, unchanged, and says so. Email takes the same file to 1.7 MB.
It is a 1-bit scan. Black and white scans (formats CCITTFaxDecode or JBIG2Decode in the inspector) are already stored very efficiently. The compressor leaves them untouched.

Maps, plans and CAD exports: the weight is drawings

A site plan or a map exported from a drawing program can be large with no pictures at all: the weight is thousands of vector paths. Image settings cannot touch it, and our compressor keeps drawings exactly as they are, so lines stay sharp at any zoom. Re-exporting from the source program with fewer layers or a simpler level of detail is the real fix; after the fact, removing pages is.

What still makes it smaller

These work on any file, because they change how much document there is, not how it is encoded.

Send only the pages they need. If the receiver needs 5 pages of 60, extract those pages into a new file. Nothing beats it.
Delete the pages nobody needs. Cover sheets, blank backs of scans, appendices: delete them and keep the rest.
Split it into parts under the cap. Cut at section breaks so each part stands alone, and send them as numbered files:

Split PDFBy ranges, every N pages, by bookmarks or size. Free, runs on your device.

Grayscale a colour scan. On a colour scan of black on white paper, grayscale alone took our 10-page test scan from 17.6 MB to 15.4 MB. Once the Email preset has run, grayscale adds little, so use it when you also want pages that print predictably.

When the file is as small as it truthfully gets

Some files have no slack left. A long text document, a 1-bit scan, a technical drawing: their size is the content. And one kind of file should not be compressed even when it could be: a PDF with a certificate signature, which any change invalidates. The inspector above shows it under Digital signatures; compressing a signed PDF explains how to tell and what to do instead. Then the remaining options are about sharing, not squeezing: a link instead of an attachment (the guide on sending a large PDF covers the trade-offs), fewer pages, or asking the receiver whether their cap really applies. Any tool that promises a fixed percentage on every file is describing a picture-heavy file, not yours. How small can a PDF get has realistic floors by file type.

The inspector and the compressor both run on your device; nothing is uploaded while you find out.

Questions

Why is my PDF the same size after compressing?

Because its size is not in pictures a compressor can shrink. It is text, fonts or vector drawings, or the pictures were already compressed. Our compressor then returns the original unchanged and says why, instead of handing back a copy that is no smaller.

Will a different compressor do better?

On pictures, a little, by going to a lower resolution. On text, fonts and drawings, no tool can do much without changing the document. If the inspector says pictures are a small share of the file, stop trying compressors and remove pages instead.

Does compressing twice help?

Rarely. The second pass finds pictures that are already compact and leaves them alone, or re-encodes them again and stacks JPEG damage. Compress once from the original, at the setting you need.

I picked the Print preset and nothing changed. Why?

Print keeps pictures at 300 dpi. A scan made at 300 dpi is already there, so there is nothing to shrink and the tool returns it unchanged. Pick Email or Web, or type a size target.