100% on-device
Compress a Word document
A .docx that is too big to email is almost never too much text — it is the screenshots pasted into it. This shrinks those and leaves your document alone.
How it works
- Drop your .docx. It is unpacked in this tab — a .docx is a ZIP of XML plus your images.
- Pick a target. Email (1024px) · Balanced (1600px) · Sharp (2200px) — the ceiling for every image in the document.
- Download. Images re-encoded, document re-zipped, text untouched.
In our own test, a 1.8 MB document came out at 69 KB — 96% smaller.
word/media is most of a heavy .docx, and it is not the only passenger
Text is close to a byte per character and compresses again inside the ZIP, so prose is never the reason a document will not send. Beyond the images, three things quietly add weight: a pasted Excel chart embeds the entire source workbook as an object, drawings pasted from other Office apps arrive as EMF metafiles, and revision identifiers accumulate on every single save.
Only the first of those is an image problem. If a document has been through dozens of rounds of tracked changes, or carries a chart that is really a workbook, the weight sits outside word/media and no picture pass will find it. Copying the text into a fresh document is the blunt fix for that case.
A pasted screenshot is stored at capture size, not at the size on the page
Dragging a picture’s corner in Word changes a scale attribute in the XML; the stored pixels do not change at all. A screenshot taken on a high-density display arrives at twice the linear resolution it was viewed at, and is then shown four inches wide while every one of those pixels stays in the file at full count.
PNG makes it worse than it needs to be. Its filters depend on neighbouring pixels repeating, which holds for flat panels and stops holding once you add antialiased text, gradients and drop shadows. Inserting an image from a file you have already sized, rather than pasting it from the clipboard, avoids the whole problem before it reaches the document.
Questions
Is my document really not uploaded?
Yes — and not as a policy, as a fact of how it is built. There is no upload endpoint: the file is read straight into this browser tab and compressed by code running on your own machine. You can watch the Network tab while it works, or pull your Wi-Fi out mid-compression and it will still finish. Nothing is stored, nothing is queued, and there is no account.
Will my formatting survive?
Completely — the tool never parses your document XML. It reaches only into the media folder inside the .docx, recompresses the images, and repacks. Text, styles, headers, tables, footnotes, tracked changes and comments come out byte-for-byte identical, because nothing in that machinery is ever read, let alone rewritten.
Nothing happened — it says there is nothing to shrink.
Then your document really is just text, and it is already small. Word's XML is compressed inside the ZIP already; there is no honest way to make a text-only .docx meaningfully smaller, so the tool tells you that rather than shaving a few bytes and calling it a result.
Why are pasted screenshots so heavy?
Word stores a pasted screenshot as a PNG — a lossless format built for sharp edges, not for the thousands of subtly different colours in a real screen. That single paste can be several megabytes. When such an image has no transparency, this tool converts it to JPEG, which is usually where almost all of the savings come from.
What about old .doc files?
The pre-2007 .doc format is not a ZIP and cannot be read this way. Open it in Word, "Save As" .docx, then compress.
Other formats
The same workbench handles all of them — pick a tab above, or start from its own page: