ZIP Extractor

Open a ZIP and take files back out, or pack files into a new one.

Create a ZIPOpen a ZIP

Everything happens in this tab. The archive is never uploaded, and neither are the files you take out of it.

Drop a .zip file here to see what is inside

Your file never leaves your device · No signup · No daily limit

Up to 256 MB per file

This device opens archives up to 256.00 MB.

How it works

Two modes share the page: Open a ZIP and Create a ZIP. To open one, drop a.zip file. The archive is read into memory whole, so there is a ceiling that depends on your device (256 MB on a desktop, 64 MB on a phone or a low-memory laptop) and the figure that applies to yours is printed above the drop zone.
Browse the contents. Names, compressed sizes and uncompressed sizes come from the archive's own central directory, before a single byte is decompressed.
If the archive is encrypted, the file names are usually still readable, because the index itself is not encrypted. The contents are not. This tool decrypts neither ZipCrypto nor AES, which is why it never asks for a password. What it does do is name the protection in use, so you know which desktop tool to reach for.
If old filenames look like gibberish, switch the character set. The list re-reads instantly, because nothing has been decompressed yet.
Save one file or all of them. Entries are inflated one at a time and written straight out as downloads, so only one is ever in memory. In the other direction, Create a ZIP takes the files you drop and a compression level and packs them into a single archive.

What people use it for

Opening a ZIP on a machine where you cannot install software.
Checking what is inside an archive someone sent you before you extract it anywhere.
Pulling one file out of a large archive without unpacking the rest.
Opening an archive that holds customer data, configuration or key files. The kind that should not be handed to an upload site.
Reading an old archive whose Turkish, Japanese or Cyrillic filenames come out as mojibake in other tools.
Packing a handful of files into one ZIP to attach or hand over, without installing anything.

Good to know

Password-protected archives cannot be decrypted here. Names, sizes and the protection in use are listed (the ZIP index is not encrypted, which is why they are readable) but the contents are not opened. ZipCrypto and AES-128, 192 and 256 are recognised and named; none of them is decrypted. If you know the password, 7-Zip, Keka, WinRAR and Windows Explorer will open the archive with it.
There is no password field, because there is nothing this page could do with what you typed into it. Asking for a password it cannot use would be worse than saying up front that it cannot decrypt. In a mixed archive the unencrypted entries still save normally; the locked ones are badged and refused.
It cannot recover, crack or bypass a password you do not have, and no honest tool can. A readable file list is not a sign the archive is open: names inside an encrypted ZIP are usually readable without the password, the contents are not.
PKWARE Strong Encryption encrypts the file list as well, so in such an archive even the names may not read correctly. It is rare, and it gets its own named explanation rather than a generic failure.
There is a size ceiling, and it is stated before you drop anything, because the archive is held in memory whole for the whole session rather than streamed from disk. Desktop: 256 MB per archive, 192 MB for a single unpacked entry, 192 MB for one "save everything" run. Phones and machines with 4 GB or less: 64 MB, 48 MB and 48 MB. An entry over the ceiling is refused before it is attempted, with its real size and the limit side by side. The ceilings exist because a tab that dies mid-extract takes the work with it and says nothing.
Creating a ZIP is memory-bound too. Measured peak heap was 2.95x the input bytes at the balanced level and 1.91x when storing without compression, so packing input is capped at 160 MB on a desktop and 40 MB on a low-memory device. Speed was measured as well: about 41 ms per MiB at the balanced level.
"Smallest" is only 0.1–1.3% smaller than "Balanced" for roughly twice the time, measured here, and files that are already compressed (JPEG, PNG, MP4, PDF, ZIP and the like) are stored untouched at every level, because deflating them again gains nothing. That is why there is no maximum-compression slider that quietly does nothing.
ZIP64 is read properly: sizes over 4 GB and archives with more than 65535 entries are reported from the 64-bit index, so the numbers in the list are correct. Worth stating because browser-based ZIP tools frequently get this silently wrong. Reporting a size correctly is not the same as being able to open the archive. It still has to fit the device ceiling above. The list renders the first 500 entries; "Save every file" still covers all of them.
Filenames written by modern tools are tagged as UTF-8 and come out right in every language. Older archives (DOS-era, older Windows Compressed Folders, some WinRAR output) record no encoding at all, so their names have to be guessed. CP437 is the default, with a picker for CP850, CP857, UTF-8, Shift_JIS, EUC-KR, GBK, Big5, IBM866 and Windows-1250/1251/1252/1254. Chromium's TextDecoder rejects the three DOS code pages by label, so those tables ship inside the tool rather than being borrowed from the browser.
That guess cannot always be right: two code pages can turn the same bytes into two equally plausible names. That is why you get a choice and a live preview instead of a promise that detection is automatic. A � in a name means the selected character set has no character for that byte. The name is wrong, not merely unusual.
ZIP only, on purpose. 7z, RAR, TAR, GZ and ISO are not opened. Inside a ZIP, only Store and Deflate can be inflated; Deflate64, BZip2, LZMA, Zstandard, XZ and PPMd are listed by name in the contents but cannot be extracted, and you get a message naming the method and pointing at 7-Zip, Keka or WinRAR.
Some archives contain paths that try to escape their folder, ".." segments or absolute paths. The stored path is shown exactly as it is and flagged, and it is never used when saving. A tool that quietly rewrites such a name is showing you a filename the archive does not actually contain.
A small archive can declare a huge payload. Both sizes are read from the index before anything is inflated, so a suspicious ratio (unpacking to 1 MB or more at 100 times its stored size) is flagged first and only unpacked when you ask. And because that declared size is written by whoever made the file, extraction is also stopped the moment an entry writes past what it declared. We tested that against an archive declaring 1024 bytes that really inflates to 64 MiB: it stopped after 45 ms with 6 MB of heap in use.
These figures were measured in desktop Chromium 151. Firefox, Safari and phone browsers were not measured (Safari's per-tab memory ceiling is materially lower) so read them as desktop Chromium numbers rather than a promise for every browser.

Frequently asked questions

Can it open a password-protected ZIP?

It can list what is inside one (names, sizes and which protection is used) because a ZIP's index is not encrypted. It cannot open the contents: neither ZipCrypto nor AES is decrypted here, which is why you are never asked for a password. If you know the password, 7-Zip, Keka, WinRAR and Windows Explorer will open the archive with it.

Why is there no password field?

Because there is nothing this page could do with the password you typed. The engine that inflates bytes here decrypts neither ZipCrypto nor AES, and a field that is guaranteed to fail is less honest than saying so up front. No password field also means there is nowhere for a password to go.

Why are the filenames gibberish?

The archive is old enough that it does not record which character set its names use, so they have to be guessed. Pick the encoding that matches where the file came from: the list re-reads instantly, because names come from the index and nothing is decompressed to show them.

How big an archive can I open?

It depends on the device, because the archive is held in memory whole rather than streamed: 256 MB per archive and 192 MB per unpacked entry on a desktop, 64 MB and 48 MB on a phone or low-memory laptop. The figure for your device is printed on the page, and an entry over the limit is refused before it is attempted, with the reason.

Does it open RAR or 7z?

No. ZIP only. Supporting 7z or RAR would mean shipping a multi-megabyte decoder for formats this page does not promise. Inside a ZIP, only Store and Deflate are extracted; BZip2, LZMA, Zstandard and PPMd entries are listed but cannot be unpacked.

Is my archive uploaded?

No. Your browser opens it from your disk; no part of the archive, and no extracted file, is sent anywhere. People routinely hand archives full of customer data, configs and key files to upload sites. That is exactly the transaction this page avoids.

Related tools

Percentage Calculator