Extract a ZIP archive

Most tools that open an archive hand you everything at once. That is fine for three files and tedious for a hundred, and it is actively unhelpful when somebody has sent you a 400 MB archive because you needed one spreadsheet out of it.

The listing here comes first: every file, its name and its size. Download the one you want and leave the rest, which on a phone or a metered connection is the difference between a useful tool and an expensive one.

ZIP works in the browser because the browser already has what it needs — the deflate decompressor is built in, and the archive structure is a header, the bodies and a directory at the end. Nothing has to be downloaded to read one.

RAR, 7Z, TAR.GZ, BZ2 and XZ are different. Each needs a decoder the browser does not have, and shipping several megabytes of decompressor for a format most visitors will never drop is the wrong trade. Those go through the conversion pages instead, which turn them into ZIP on our server — and the file is identified by its contents, so if you drop one here you are told what it actually is rather than given a vague failure.

An encrypted ZIP is refused for a different reason: without the password there is nothing to read. The same is true of an archive that uses LZMA inside a ZIP container, which is legal and rare and not something the browser can decompress.

A password-protected ZIP comes in two varieties. The original scheme, added with PKZIP 2.0 in 1993, is broken and recoverable without the password; AES-256, added later, is genuinely strong. Both present the same password prompt and nothing on the outside says which you are facing.

The practical consequence is about compatibility rather than security: an AES-encrypted ZIP cannot be opened by Windows Explorer, which still only understands the old scheme. If an archive refuses a password that is definitely correct, an encryption mismatch is a more likely explanation than a typo.

A ZIP can be spread across numbered volumes — `.z01`, `.z02` and a final `.zip` — to get past a size limit. They are not independent archives: every part must sit in the same folder, and extraction starts from the `.zip` rather than from the first numbered file.

A missing or renamed volume usually makes everything after it unrecoverable, and the failure looks like corruption rather than an absence. If an extraction stops partway with no clear reason, checking that every numbered part arrived is the first step.

ZIP had no standard way to record the character encoding of file names for most of its history, so an archive created on a Windows machine in one region and opened in another shows accented and non-Latin names as nonsense. The files themselves are perfectly intact.

Modern tools write UTF-8 and set a flag saying so, which resolves it. Older archives frequently do not, and there is no way to recover the intended name with certainty — only to guess the code page. It is the most common reason an extracted folder looks damaged when nothing is.

Extraction runs inside your browser tab, so the archive is never transmitted. That is worth more here than for most operations, because an archive is by definition a container of everything somebody bundled together — a backup, a client folder, a set of scanned documents.

There is also no size limit beyond your own memory and no waiting for an upload, which is the difference between extracting a multi-gigabyte archive and being told it is too large to accept.