How to create a ZIP file
- Drop files or whole folders into the box above, or use Choose files and Add a folder. You can add more in several goes and remove any entry from the list.
- Name the archive and pick a compression level. The defaults suit most jobs.
- Press Create ZIP. When it finishes, the page lists every file with its original and packed size, and tells you whether the archive checked out.
- Download the ZIP. Nothing was uploaded — the archive was built in your browser tab and exists only on your device.
What "checked" means here
Your files never leave your device, and that part is easy to promise. What matters just as much is whether the archive opens on the other side, because a damaged ZIP is discovered by whoever receives it. So after building the file this page does two things:
- It reads the archive's own central directory — the index at the end of every ZIP — and takes every name, size and method in the table from there, not from what it intended to write.
- It unpacks the finished archive in memory and compares each file with its source by CRC-32. A file that does not match is named and the result is marked as unusable.
Above 256 MB of source files the second step is skipped, because unpacking that much in a browser tab would risk running out of memory; the page tells you when that happens.
Stored or compressed?
A ZIP can compress a file with deflate or simply store it. Text, code, CSV, SVG and uncompressed images such as BMP shrink a lot — often by two thirds or more. JPEG, MP4, PDF, Office documents and other ZIP files are already compressed, and running deflate over them takes time and saves almost nothing. With Store already-compressed files on, those types go in as they are and the table marks them Stored. Untick it if you want everything deflated regardless.
| Level | Use it when |
|---|---|
| Fast (1) | Large folders of text where speed matters more than the last few percent |
| Balanced (6) | Almost everything — the setting most ZIP tools default to |
| Smallest (9) | An attachment size limit is close and you can wait a little longer |
Names, folders and duplicates
Folder structure is kept by default, so a folder you drop in comes out as the same folder when the ZIP is extracted. Untick Keep folder structure to put every file at the top level.
Flattening or merging can leave two files with the same name. A ZIP may technically hold
both, but the second overwrites the first when it is extracted — quietly, on most
systems. Names that repeat, ignoring capital letters because Windows and macOS treat
Report.pdf and report.pdf as one file, are numbered
(report-1.pdf) and listed in the report. Names with accents or non-Latin
characters are flagged as UTF-8 inside the archive, and each file keeps its last-modified
date.
Limits worth knowing
- No password. The classic ZIP password scheme is weak enough to be broken quickly, and strong AES encryption needs a different library, so this page does not offer one rather than offer a false sense of safety. To protect a file, encrypt it with a tool built for that.
- 4 GB. The archive format used here has a 4 GB ceiling per file and per archive. Anything larger is refused with a message instead of producing a corrupt file.
- Memory. Files are read in a stream and the archive is assembled piece by piece, but a very large job still depends on your device having room.
Next steps
Sending a PDF? Compress it or merge several first. Have images that are too big? Compress them to a target size before zipping. Every converter is listed on the data tools page.