pdfviz

Compress PDF
to 20 KB

or drop one here
Compression level
Target size
Choose a PDF file first.

20 KB is 20,480 bytes, and almost nothing with a photograph on it reaches that. A short text document can. A scanned page, in any realistic quality, cannot. This page says so up front so you do not spend twenty minutes finding out, and shows what to aim for instead when the form demands a number the document cannot meet.

How to compress a PDF to 20 KB

  1. Choose the PDF

    Press Choose a PDF, or drop the file on the page. The file is read in this browser tab and is not sent to any server.

  2. The target is already set

    This page opens with the target fixed at 20 KB, so there is nothing to type. Pick a different compression level if you want, or enter your own larger number in KB if the document needs more room.

  3. Press Compress

    Four methods race against each other: copy the pages into a fresh file, re-encode the placed images, cut each embedded font down to the glyphs the pages draw, and only when no other method reached the target, redraw the pages as images with the text kept on top.

  4. Read the check, then download

    Every candidate is opened again and its characters are counted against the source. The page shows the size, the method that won, and the characters kept. If no method reached 20 KB, it ships the smallest verified candidate and states the real floor.

The four methods are described in full, with the measurements behind them, in how it works. The whole process runs in your browser through WebAssembly, which means your file is never uploaded and no account is needed.

Why 20 KB is out of reach for scanned PDFs

20,480 bytes is smaller than a single low quality thumbnail. A scanned A4 page compressed to that size loses the letterforms, not just the sharpness. There is no setting and no tool that changes this, because the information is simply not representable in that many bytes.

The hardest target the public bench measured was 100 KB, five times this: a 20-page text file went from 124.2 KB to 89.2 KB in all three runs. Nothing in that run suggests 20 KB is reachable for a document with page images in it. A one page scan at 200 DPI starts at roughly 500 KB, and even at 72 DPI it sits well above 20 KB.

If a form is genuinely fixed at 20 KB and your document is a scan, the form is almost certainly asking for a JPEG or PNG image rather than a PDF. It is worth reading the instructions again before you spend time on a target the format cannot meet.

What documents do reach 20 KB

A one page text document with a single font, once the font is subset to the glyphs it draws. A signature or a logo on a white page. A generated receipt or ticket with no photograph. These are genuinely small documents, and the compress tool gets them there.

Font subsetting is what makes it possible. An embedded font ships with every glyph it was designed with, often thousands, and a one page document draws a few dozen. Cutting the font down to what is used is often the difference between 200 KB and 20 KB on a text file. The structural rebuild also helps, because it drops orphaned objects and old revisions that accumulate every time a file is saved.

A plain text export from a word processor is the easiest case. Remove the images, save as PDF, and let the font subset do the work. The output renders as sharp as the original at any zoom, because no pixel was touched.

What to aim for instead of 20 KB

Run the compression anyway and read the number it reached. That tells you the floor for your document. If the smallest proven file is 46 KB, then 46 KB is the answer, and the form is asking for something that does not exist.

A 50 KB target is the smallest that a simple scan sometimes reaches, and a 100 KB target is the smallest that is reasonable to plan around. Both use the same engine and the same verification checks as this page.

If the form is fixed at 20 KB and your document is a scan, consider exporting the page as a JPEG instead. PDF to JPG can produce a small image file that may fit the cap where a PDF cannot.

Real use cases for a 20 KB target

Generated receipts and tickets with no photographs. A one page text document with a single font subsets down to a few thousand bytes once the unused glyphs are stripped.

Signature pages where a scanned signature sits on a white background. The white compresses to almost nothing, and the font subset removes any embedded text fonts.

Plain text exports from a word processor, saved as PDF without any images. These are the documents where 20 KB is a realistic target rather than an impossible one.

Questions about compressing a PDF to 20 KB

Can a scanned PDF be compressed to 20 KB?
Realistically, no. 20,480 bytes cannot hold a readable page image, and no tool changes that. A text only document can reach it once its fonts are subset, but a scan cannot.
What is the smallest file the tool produced for my document?
Whatever it reports after the run. When no method reaches the target, the engine ships the smallest candidate that passed verification and states the size. That number is the real floor for that document.
Should I lower the scan resolution to hit 20 KB?
You can, and it will still not be enough for a full page. It also destroys the print quality of the thing you are sending. If the requirement is genuinely 20 KB, the form almost certainly wants a small image rather than a document.
Why does font subsetting help so much at tiny sizes?
Because a font is a program carrying thousands of glyph outlines, and the page uses a few dozen. On a short text document the fonts can be most of the file, so cutting them to what is drawn removes most of the weight in one step.
Is 20 KB enough for a one page text document?
Often yes, if the document uses a single font and carries no images. The font subset plus the structural rebuild can bring a short text PDF well under 20 KB without touching a pixel.
What should I do if the form demands 20 KB and my scan cannot reach it?
Export the page as a JPEG with PDF to JPG and check whether the form accepts image files. A small JPEG can fit where a PDF cannot.

The other pdfviz tools

They all run in a browser tab like this one. Each reads its own output back and checks it before offering the download, and none of them needs an account.

Where to go next