Create ZIP

Pick any files — or drop in a whole folder — and get a single .zip that opens anywhere. The archive is built on your own device, so nothing is uploaded.

Drop files here or choose files
Pick several files at once, or drag in a whole folder · files are not uploaded · 50 MB per file, 100 MB in total
.zip

The level sets how hard the compression works, not how many files you can add: 0 doesn't compress and is the fastest, 6 balances size against speed and is recommended, 9 uses the strongest compression setting and may take longer, depending on the files and device. Use 0–1 for images, video and files that are already compressed, and 6–9 for text, logs and source code. This tool offers 0 to 9, with 6 as the default and 9 as the hardest setting it will compress at.

When ticked, files are written into the archive at their relative path, like docs/2026/report.xlsx (your browser only reports a path when you drag in a whole folder). When unticked, every file sits in the root of the archive and a second file of the same name becomes "report (1).xlsx" instead of overwriting the first.

🔒 Files are zipped inside your browser and never uploaded; the zip comes without a password.

How to use

  1. Add the files: Drop your files onto the box or click to choose them. Any file type is accepted, you can add more in several goes, and on a computer you can drag in a whole folder. Each file can be up to 50 MB and the selection up to 100 MB in total.
  2. Name the archive and set the options: Type a name in Archive name (left empty, the download is called archive.zip). Leave Compression level at 6 unless you know you need otherwise, and untick Keep folder structure if you want every file in the root of the zip.
  3. Create the zip and save it: Click Create zip. You'll see how many files went in and the original total next to the zip size, then download the archive. On a phone the button reads Save to device: if your browser can share files it opens the system share sheet, so the zip goes straight into Files or an email, and if it can't, the zip simply downloads.

How this zip tool works

The tool builds one standard .zip archive from the files you pick, and it does the whole job inside the tab you already have open. Your files are read from your disk into memory, compressed with fflate — a JavaScript implementation of DEFLATE, the compression method built into the zip format — and handed back as a download link your own browser creates. There is no upload step anywhere in that chain, so contracts, payslips, ID scans and client work never leave your device.

What comes out is an ordinary zip, not a special format. Windows File Explorer, the macOS Archive Utility, 7-Zip, WinRAR and the built-in extractors on Android and iOS all open it directly, so the person receiving it needs nothing installed.

Big files are handed to a background thread (a Web Worker) to be compressed, which keeps the page responsive on the kind of job that would otherwise lock it up. Smaller files — under roughly 160 KB each, which is how fflate splits the work — are compressed on the page's own thread instead, so a batch of several hundred small files can still make the page stutter for a moment. The button reads Zipping… and is disabled until the archive is finished. Nothing is kept afterwards: refresh or close the tab and both your selection and the finished archive are gone. No localStorage, no cookie, no server copy.

Which compression level should I use?

0  store only — no compression, archive ends up ~100 bytes per file larger
1  light — CSS sample 189,614 B → 44,070 B
6  default — same sample → 34,813 B (82% smaller)
9  hardest — same sample → 34,788 B (25 bytes better than level 6)

Leave the slider at 6. On our three test files, level 9 saved 25 bytes, 0 bytes and 0 bytes respectively compared with level 6, for the same archive in every practical sense — and it asks the device for more work to get there, so it can take longer, depending on your files and your hardware. We measured the sizes, not the time. The slider offers 0 to 9, 6 is the default, and 9 is the hardest setting this tool will compress at; anything you type outside that range is pulled back to 0 or 9.

Move the slider down to 0 or 1 when everything you are packing is already compressed — photos, video, music, existing .zip or .rar files, or Office documents, which are themselves zip containers. There is usually little room left for DEFLATE to squeeze those further, and how little depends on the individual file, so you are mostly paying in time. Level 0 stores the bytes as they are, which is the least work you can ask of the device.

Push it up to 9 only when the batch is text: source code, CSV exports, log files, SQL dumps, SVG. Those have the kind of repetition DEFLATE feeds on, and on a slow upload link the extra packing time can be worth a few more percent.

Why a zip isn't always smaller

How much you save depends entirely on what you put in, and with already-compressed files the honest answer is "almost nothing". Zip compression is lossless — it only removes repetition in the data, never quality — so the savings come from files that contain repetition in the first place. A 189,614-byte CSS file came out at 34,813 bytes at the default level, 82% smaller. A 73,092-byte PNG image came out at 71,006 bytes: 3% smaller. A small .zip we tried to re-pack grew from 652 to 752 bytes at level 0, because every entry in an archive carries about 100 bytes of header.

That is the normal outcome for JPEG, PNG, GIF, WebP, MP4, MP3 and for .docx, .xlsx and .pptx files, which are zip archives wearing a different extension. With those, zipping is still worth doing — it just isn't about space. One attachment instead of thirty is easier to send, survives email clients that mangle long attachment lists, and arrives with your folder layout intact.

One more thing worth knowing: zip compresses each entry on its own, with no de-duplication across files. We packed the same 9,249-byte source file 40 times over — 369,960 bytes in — and got 145,922 bytes, a 60.6% saving, which is almost exactly the 60.4% the single file manages by itself. Forty identical copies cost forty times as much as one. Cross-file de-duplication needs a solid archive (7-Zip's .7z), which this tool doesn't produce.

Folder structure, file names and the archive name

Keep folder structure is ticked by default and writes each file into the zip at its relative path — docs/2026/report.xlsx — so whoever extracts it sees the same tree you had. Folders are only kept when your browser hands the tool a relative path for each file, which in practice means dragging a whole folder in on a computer. On a phone you can usually only pick files one by one rather than a whole folder, so an archive made there normally comes out flat. You can change the tick right up to the moment you click Create zip; it is applied when packing, not when a file is added. If your browser won't read part of a folder you dropped — an item it refuses to open, or a sub-folder it stops listing — the tool names what it couldn't read and adds nothing from that drop, because half a folder shouldn't pass for the whole one; drop it again, or choose the files yourself.

Untick it and every file lands in the root of the archive under its plain name. Duplicate names are handled rather than lost: a second report.xlsx becomes report (1).xlsx, so nothing silently overwrites anything else. Names with accents or non-Latin scripts — Vietnamese tone marks, Spanish tildes, Chinese or Arabic characters — are written as UTF-8 with the zip language-encoding flag set, which is what stops them turning into gibberish in other people's extractors. One name is turned away rather than packed: a file called exactly __proto__, with no extension after it. The library that builds the archive files its entries under their names internally, and that one name means something else there, so the file would quietly go missing from the zip — the tool stops and asks you to rename it first. __proto__.txt, or a __proto__ sitting inside a folder, packs like anything else.

Whatever you type in Archive name is tidied before use: surrounding spaces go, a trailing .zip is dropped so you don't end up with report.zip.zip, the characters Windows and macOS reject in a file name (\ / : * ? " < > |) become hyphens, and anything past 80 characters is cut. Leave the box empty and you get archive.zip.

Occasionally a dragged-in folder produces a path with ../ in it or a drive letter such as C:\ — paths that would let an archive write files outside the folder someone extracts it into. The tool refuses the whole batch in that case and tells you to untick Keep folder structure, which flattens everything to plain names and packs normally.

What this tool can't do

It can't put a password on the archive. The zip you download is unencrypted, by design — neither ZipCrypto nor AES is implemented here. If you need encryption, make the archive locally with 7-Zip and choose AES-256. Worth knowing before you rely on it: an encrypted zip hides the contents of the files, but in most tools the list of file names inside stays readable.

Only .zip comes out. There's no .rar (licensing), and no .7z, .tar or .tar.gz. You also can't add files to an archive you already have, or rename and delete entries inside one — each run builds a fresh archive from the current selection. Multi-part archives (.z01, .zip.001) aren't supported either, so a batch that's too big for the recipient's limit has to be split into separate runs by hand.

A zip from here carries names and bytes, nothing else. Modification dates are set to the moment you pack, not the dates on your originals; permission bits, ownership and macOS extended attributes and resource forks are dropped; and empty folders disappear, because a browser hands the page a list of files and a folder with nothing in it never appears in that list. File contents themselves are exact — any level, including 0 and 9, extracts byte for byte identical to what you put in.

The size ceilings are 50 MB for any single file and 100 MB for the selection as a whole; go over the total and you'll be asked to remove some files before packing. There is no limit on how many files you add, only that total. Both limits exist because everything is held in your device's memory while the archive is built, which is also why a phone will struggle closer to 100 MB than a laptop will — split a large job into two or three archives instead.

Frequently asked questions

Are my files uploaded to a server?
No. The files are read and zipped inside your browser, and the download link points at an archive that was built on your device — there is no upload at any stage. Nothing is saved either: close or refresh the tab and both your selection and the archive are gone.
How many files can I zip at once?
There's no limit on the number of files. The limits are 50 MB for any single file and 100 MB for everything you've selected together, because the whole batch is held in your device's memory while the zip is built. Over the total and the tool asks you to remove some files first.
Can I add a password to the zip?
No — the archive is created without encryption, and there's no option to add one. For a password-protected archive, use 7-Zip on your own computer and pick AES-256. Note that encrypted zips usually still show the list of file names inside, even before the password is entered.
Which compression level is best?
Start with 6, the default. Level 9 beat it by 25 bytes on one test file and by nothing at all on two others, and it asks the device for more computation to get there, so it can take longer — how much longer depends on your files and your device, which we didn't measure. Drop to 0 or 1 for photos, video and existing archives, and raise it to 9 only for text-heavy batches like source code, CSV or logs.
Why is my zip barely smaller than the original files?
Because those files were already compressed. JPEG, PNG, MP4, MP3 and .docx/.xlsx/.pptx (which are zip containers themselves) leave little repetition for zip to remove, and how little varies with the file — a PNG in our tests shrank by 3%. Text files behave very differently: a CSS file came out 82% smaller at the default level.
Does it keep my folders, and does that work on a phone?
Folders are kept when your browser reports a relative path for each file and Keep folder structure is ticked — in practice, when you drag the whole folder in on a computer. On a phone you can usually only pick files one by one rather than a whole folder, so a zip made there normally comes out flat. Duplicate names get a (1) suffix rather than overwriting each other.
Can I make a RAR or 7z file, or add files to an existing zip?
No to both. The tool only writes .zip — no .rar, .7z, .tar or .tar.gz — and every run creates a brand-new archive, so you can't append to, rename inside, or edit an archive you already have. Split archives such as .z01 aren't supported either.