How it works : Zip and unzip

No. Open the inspector with the F12 key, go to the “Network” tab, clear the list, then drop your files in and build the archive: nothing leaves. Cut your connection and do it again, the tool still works. It is the most important point on an archiving tool, because what people put into one is generally everything they were about to send.

Because your files are probably already compressed. A JPEG photograph, a PDF, a video, a modern office document: all of them already contain compression, and there is nothing left to gain on them. The tool tells you line by line, and it stores those files as they are rather than making them BIGGER, which is what compressing them a second time would do. What compresses well is text, older spreadsheets, code files, lossless images.

Having written the archive, the tool reopens it with its own reader, as another piece of software would: it finds the directory, walks the entries, decompresses each one and recalculates its check value from the bytes obtained. An archive can have an impeccable structure and a wrong check value: it would pass every structural test and fail at your recipient's end. It is the same requirement as everywhere on this site: we read the file we produced, we do not take the tool's word for it.

Because a browser sends down files, not directory trees: it has no way to create a folder on your device, and that is a good thing. An entry filed inside a subfolder therefore comes out with its path flattened into its name, which keeps the information without inventing a structure. That is also what makes an old archive trap harmless: an entry whose path climbs out of the extraction folder cannot write anything elsewhere, and we clean its name regardless.

Because a site triggering several downloads from one click is behaviour browsers are suspicious of, rightly. The “download multiple files” permission is asked once, and granting it is riskless here since the files come out of your own archive. We space the sends three hundred and fifty milliseconds apart, which stops them cancelling each other, and you can always extract one entry at a time.

We do not plan to, and for a reason of substance. The historic encryption of zip archives has long been considered broken; the modern one is a proprietary extension of the format, with several variants and differences between programs. So we would do heavy work to deliver a protection we could not honestly call strong. The encryption tool in the security corner solves the real problem: it shows its whole method, it uses six hundred thousand key-derivation rounds, and the file it produces is decrypted on the same page.

Not directly: the tool opens an archive or builds one, it does not modify one. The way through is to extract what you want, then drop it all back in together with the extra file. That is one more gesture, but it avoids a whole class of nasty surprises: an archive edited in place can keep traces of its previous contents, and a tool promising to “update” an archive without rewriting it entirely is not as clean as it looks.