Common questions

Why is my Excel file so large?

An .xlsx is a ZIP holding XML, images and other parts, so its size is the sum of those parts. The usual culprits are the worksheet data itself, pivot caches, embedded images, and formatting applied thousands of rows past the last real one. This tool lists every part and what it costs, so you can see which one is responsible instead of guessing.

Does shrinking change my formulas, values or formatting?

Not with the reductions that start on. They keep values, formulas, formatting and every object, and remove only what repeats what the file already says: empty cells whose format their row or column supplies, a second copy of an identical picture, or the calculation chain Excel rebuilds. The reductions that remove something start off, and run only if you switch them on under Advanced settings on the front page, or in the desktop app. Each one has a page that says what it removes. With every reduction off, only the compression changes and every part stays byte for byte identical. The result is checked against the original before a download is offered.

How much smaller will my file get?

It depends on what is inside. In the 16-workbook run behind the home page's figures, with every reduction on, which is Maximum here, the engine saved 37.5% of the total size. Those workbooks were chosen so every reduction shows a saving, so they come out better than a typical folder would. The highest reduction we have measured is 99.9%: a 10.4 MB workbook formatted far past its data came out at 6 KB. These test files do not predict your result. The tool checks your workbook and reports its actual size saving before offering a download.

Is my spreadsheet uploaded to a server?

No. There is no upload and no server involved in processing. The page is static files and the work runs in WebAssembly inside your browser tab. You can confirm it by opening the Network tab in your developer tools while you shrink a file, or by disconnecting from the internet and shrinking one anyway.

Does it work with .xls, .xlsm or .xlsb files?

Macro-enabled .xlsm workbooks and .xltx or .xltm templates work; the macros pass through untouched. Legacy .xls and binary .xlsb are different formats and are refused with a reason, as are encrypted and digitally signed packages. Open an .xls in Excel and save it as .xlsx first.

Will a smaller file open or calculate faster in Excel?

It depends on what made it big. Recompression only changes how the file is packed, so Excel opens it at the same speed. Formatting that runs far past your data and thousands of hidden shapes are different. Excel reads them every time it opens the file, and Microsoft notes that excess formatting "can slow down not just Excel, but the overall performance of your computer." Those reductions remove something, so they start off. Choose Maximum under Advanced settings to run them. Slow formulas stay as slow as they were.

How large a workbook can it handle?

Files up to 50 MB run in most browsers. Larger ones sometimes do, depending on how much memory your browser gives the page. The size on disk is only part of it. A workbook is a zip, and its XML unpacks ten times over or more, so a 30 MB file can hold 600 MB of data, and that has to fit in the tab alongside your file, a working copy and the result. Files over about 40 MB will be slower, and the page says so before it starts. For bigger workbooks, the desktop app takes up to 512 MB.

Does the desktop app have the same limits?

It has limits, and they are not the browser ones. Nothing about a workbook sets them: they are budgets we choose so a bad file cannot exhaust the machine. A browser tab is the tight case, because 32-bit WebAssembly addresses 4 GB at most and the tab dies rather than refusing, which is why the page is careful. A desktop process has real memory and can say no politely, so the app takes a workbook up to 512 MB on disk and 2 GB unpacked, and refuses one past that with a reason instead of falling over.

What does the site record when I shrink a file?

Numbers, twice. Before a run: the size of the file, how much it says it holds unpacked, and the extension. When it finishes: the size after, how long it took, which reductions ran, and where the space was. No file name, no sheet or part name, no account, no cookie. The first line exists because a workbook large enough to exhaust a tab can take the page down with it, and a run that never comes back is the one thing we cannot otherwise learn from. It is sent as a plain link you can read in the Network tab, and the privacy page lists the whole of it.

What are the reductions, and which are safe?

There are two kinds. Four keep everything and start on: they remove empty cells whose format their row or column already supplies, drop the calculation chain Excel rebuilds, keep one copy of a picture stored twice, and remove links to other workbooks that nothing in the file reads. Five remove something and start off. Advanced settings on the front page and the desktop app both list them, and each has a page that says what it removes: formatting on empty cells past the data, the saved copy of a PivotTable's source data, invisible shapes stacked on copies of themselves, the copy of values each chart saves, and links to other workbooks, broken as Excel breaks them so every value stays as it shows. On most workbooks the largest saving still comes from compressing the file again. Unknown content stays intact, and the original file is never changed.

Do I need to install anything or create an account?

No. There is no install, no sign-up, no email and no watermark. Open the page and drop a file on it. The desktop app is the paid one: it takes whole folders, runs the same engine on every workbook across every core, and writes copies into a folder you choose so every original stays where it is.

The numbers above come from 16 public workbooks: files people attached to bug reports, and large files from GitHub repositories. Drop your own file to see what it says about yours.