How the archive is read
The tool uses fflate (version 0.8), a small JavaScript library, to read the ZIP. A .zip file ends with a central directory that lists every entry, with its name, size and compression method. The tool first reads that directory itself, before inflating anything, to check the declared sizes and the encryption flag; fflate then inflates each entry. ZIP files normally store entries as either stored (no compression) or deflate (the format described in RFC 1951), and those are the two methods fflate understands. ZIP64 archives (beyond 4 GB or 65,535 files) are refused, since a browser tab will not hold data of that size.
What you see after "Open archive"
The tool lists every file with its full path and uncompressed size, sorted alphabetically, plus a total line in the form "12 files · 3.4 MB uncompressed". Folder entries are hidden and only files are listed; if the archive contains no files at all you get the generic error message. Each row has its own Download button.
Limits you should know about
| Situation | What happens |
|---|---|
| Password-protected ZIP | Not supported. The tool checks the encryption flag on every entry and stops with "encrypted entries are not supported" before extracting anything. |
| Compression other than stored/deflate (for example bzip2, LZMA or newer Deflate64) | Fails with the error message. |
| Not a ZIP (a renamed .rar or .7z, or a damaged download) | Fails with the error message. |
| Very large archive | Everything is decompressed into memory at once when you click Open archive. To protect the tab, an archive whose entries declare more than 500 MB in total is refused with a message saying how large it would be. |
| Zip bomb (a tiny file that expands enormously) | An entry of 20 MB or more that expands more than 100 times is refused. This is judged from the sizes in the central directory before anything is inflated. |
The size checks rely on the sizes the archive declares about itself, so they stop ordinary oversized or bomb-like files but are not a guarantee against a deliberately crafted archive. Open archives from sources you trust.
Downloads work differently from a desktop unzip
- The folder structure is not recreated, but paths are kept in the name. A file in a folder downloads with the folder names joined by two underscores, so
images/photo.jpgis saved asimages__photo.jpg. The list still shows the full path. - Same-named files do not collide.
a/index.htmlandb/index.htmldownload asa__index.htmlandb__index.html. Only if two flattened names are still identical is a number added at the front. - "Download all" is a series of separate downloads, one every 150 ms. Some browsers ask for permission before allowing multiple downloads, so allow it when prompted.
- It cannot create ZIP files. Extraction only.
Once a file is out, you can check its contents with the JSON formatter, convert tabular data with CSV to JSON, or compare a download against a published checksum using the hash generator, which hashes text rather than files.