Compress a PDF to 1MB, 500KB or 200KB

Pick the size you need to come in under and let the tool work out the settings. It measures your file first, uses the whole budget rather than overshooting, and tells you straight away when a number is out of reach.

Quick answer: choose your target below to open the compressor with it already set. On a ten-page 200 dpi colour scan, 1 MB is comfortably reachable and so is 200 KB; 100 KB is not, and the tool says so after one pass rather than quietly missing. Your text stays selectable throughout, and you never get back a file bigger than the one you started with.

Compress a PDF to an exact size
Nothing uploaded · text stays selectable · never returns a bigger file

2 MB 1 MB 500 KB 300 KB 200 KB 100 KB

How many scanned pages actually fit under each limit

We compressed a real 200 dpi colour scan — a page of body text, a photograph and paper grain, the kind of thing a flatbed or a phone scanner app produces — with this site's own engine, and measured the cost per page at three settings. Everything below is derived from those measurements, not from a rule of thumb.

Size limit	Sharp — 2200 px (470 KB/page)	Readable — 1325 px (148 KB/page)	Small — 450 px (20 KB/page)
25 MB	54	172	1307
18 MB	39	124	941
10 MB	21	69	523
5 MB	10	34	261
2 MB	4	13	104
1 MB	2	6	52
500 KB	1	3	25
200 KB	—	1	10
100 KB	—	—	5

Read it this way. "Sharp" keeps images at 2200 px on the longest edge, which survives zooming in on a signature. "Readable" at 1325 px is what most people would call a normal scan on screen. "Small" at 450 px is legible for body text but you will see it on a photograph. A page of pure black-on-white text costs far less than these numbers; a page that is mostly photographs costs more.

Switching on grayscale in More options takes roughly another 6% off a colour scan on top of whatever setting you chose — so a document that just misses a limit often clears it without dropping any resolution at all.

What each target does to a 10-page scan

Same document — 5490 KB, ten scanned pages — run at every target the tool offers. These are the actual outputs, including how far the images had to come down to get there.

You ask for	You get	Images capped at	Result
5 MB	4293 KB	2145 px	hit in 6 passes
2 MB	2005 KB	1544 px	hit in 6 passes
1 MB	986 KB	1052 px	hit in 6 passes
500 KB	496 KB	723 px	hit in 6 passes
300 KB	246 KB	505 px	hit in 6 passes
200 KB	196 KB	450 px	hit in 6 passes
100 KB	Not reachable — the smallest this file goes is 196 KB, reported after one pass

Two things worth noticing. The tool uses the budget rather than diving far below it: asked for 500 KB it returns 496 KB, not 150 KB, because overshooting throws away quality you asked to keep. And when a number is genuinely out of reach it says so after a single pass with the real floor attached, instead of grinding through settings to arrive at the same answer.

Why your text PDF will not get smaller — and does not need to

This is the single most common disappointment with PDF compressors, and it is not a fault in the tool. A PDF exported from Word, Google Docs or LaTeX contains text as text: characters, a font reference and some positioning. That is already about as compact as the data can be.

Word-style text PDF	Before	After	Saving
5 pages	6 KB	6 KB	no change
20 pages	23 KB	23 KB	no change
50 pages	56 KB	56 KB	no change

Fifty pages of prose is under 60 KB before anything is done to it. There is nothing to squeeze. Any tool that appears to shrink such a file is doing something else — usually rendering every page to a picture, which increases the size several times over and destroys the selectable text along with it. We measured that too: the same kind of document flattened at a middling quality came back thirteen times larger.

So if you dropped a text PDF in and it came back unchanged, that was the correct answer. This tool hands your original file back rather than returning something bigger.

When your target really is impossible

Every PDF has a floor: its text, its embedded fonts, its structure. Compression works on the pictures, so it can never go below that floor — and if your target is under it, no amount of quality reduction will help. The tool measures the floor before it starts and tells you the number.

Three things actually work at that point, in the order worth trying:

Grayscale — about 6% off a colour document, and it costs you nothing if the original was a black-and-white scan anyway.
Fewer pages — Split PDF to send only the pages that matter. Halving the page count halves the file, which no compression setting can match.
Send a link instead — for anything over about 20 MB this is what the recipient would prefer anyway.

What does not work is flattening the pages to images. On a scan it saves a little more than the sensible method and costs you the searchable text layer; on a text document it makes the file bigger. Neither trade is worth making by accident, which is why this tool never picks it for you.

Compress to a size in three steps

Open with your target set

Use one of the links above, or open Compress PDF and switch to Target file size.

Drop the PDF in

It is read in your browser and never uploaded. The tool tells you what the file is made of and what its floor is before it does anything.

Compress

You get the real before-and-after size, what changed, and either a file under your number or a straight answer about why not.

Frequently Asked Questions

Open the compressor with the 1 MB target already set, drop your PDF in and press Compress. The tool works out the image resolution and quality needed rather than making you guess at Low, Medium and High, and it searches for the gentlest settings that still meet the number — so you get the best-looking file that fits, not the first one that squeezes under. On our ten-page test scan a 1 MB target came back at 986 KB with images capped at 1052 px.

No, and the tool will tell you rather than pretending. Every PDF has a floor made of its text, embedded fonts and internal structure, and compression only works on the pictures. Our ten-page scanned test document could reach 200 KB but not 100 KB — the floor was 196 KB, reported after a single pass. If your target is below the floor the honest options are grayscale, removing pages with Split PDF, or sending a link instead.

For a 200 dpi colour scan measured with our own engine: about 2 pages at a sharp 2200 px, 6 pages at a readable 1325 px, or 51 pages at a small 450 px. A page of plain black-on-white text costs far less than any of those, and a page that is mostly photographs costs more. Grayscale takes roughly another 5 percent off a colour document on top of whichever setting you pick.

Almost certainly because it is a text document rather than a scan. A PDF exported from Word or Google Docs stores text as characters and a font reference, which is already about as compact as that data gets — fifty pages of prose is under 60 KB before anything is done to it. There is genuinely nothing to compress, so the tool returns your original file rather than a larger one.

It depends entirely on how far the target is from the file's natural size, and the tool shows you exactly what it did. Asking a 5.4 MB scan for 2 MB caps the images at 1544 px, which still looks like a normal scan. Asking the same file for 200 KB caps them at 450 px, which is legible for text but visibly soft on photographs. Because the search walks back up to the gentlest setting that fits, you never lose more quality than the target actually requires.

Yes. Target mode uses the Keep text method, which re-encodes the embedded pictures and leaves text, fonts, vector graphics, hyperlinks, bookmarks and form fields exactly as they were. A scanned document with an OCR text layer keeps that layer too.