How it works
What happens between choosing files and getting a ZIP.
The compressor is a page-level tool, not a hosted service. This page explains the path a file takes, where compression levels matter, and what the limits protect.
The short version
Read locally. Compress locally. Download locally.
- 01
Choose what goes into the archive
Drop files, pick individual files, select a folder, or load an existing ZIP. The tool reads names, sizes, and contents only after you make a selection. Folder input uses the browser’s directory picker, and dropped folders are walked entry by entry.
- 02
Let the worker build the ZIP
The page sends the selected File objects to a module Web Worker. The worker validates the job, reads each file, and adds it to a JSZip archive at the chosen DEFLATE level. Compression happens away from the main thread, so the interface can still show progress and accept a cancel request.
- 03
Review the real result, then download
When the archive is ready, the worker returns an ArrayBuffer. The page compares the finished byte count with the input, creates a Blob URL, and offers a download. Nothing is downloaded automatically. When compressing an existing ZIP, the original file remains available if the rebuild is larger.
Two modes
Create or rebuild.
Both modes end in a standard ZIP, but they start from different inputs and handle the result differently.
Create a new ZIP
Files → ZIPChoose files or a folder, set the archive name, and pick a level. Relative folder paths are kept. The input total is checked before the worker starts.
Rebuild an existing ZIP
ZIP → ZIPLoad one .zip file. The tool verifies each entry while reading, then rebuilds at the chosen level. Folder entries, dates, permissions, and comments are carried over where they are available. The output is compared with the original file.
Compression level
The trade-off is time, not format.
One slider sets ordinary DEFLATE output from Low (1) to Max (9). The level changes how hard the compressor looks for repeated data.
Low
Level 1
The quickest pass. Useful for large batches, temporary archives, or inputs that are already compressed. It still uses DEFLATE, so the file opens normally.
Balanced
Level 6
A practical middle ground between speed and size, especially when the archive contains many small files.
Max
Level 9 · Default
The slowest pass and the default. It spends more time searching for matches and tends to help most with text-heavy input. On images, video, and other compressed formats, the difference is often negligible.
A ZIP that was already created by Windows, macOS, or another ZIP tool is already DEFLATE-compressed. Rebuilding it at Max usually changes it by less than 1%; no level can restore bytes that earlier compression already removed.
Safety rails
The tab has a memory budget.
ZIP creation happens in memory. These checks run before compression so an oversized job fails quickly instead of exhausting the tab.
- Total input
- Combined size for a new ZIP, or total uncompressed content read from an existing ZIP.
- 1 GB
- Single file
- Applies to files added directly and to files extracted from an existing archive.
- 512 MB
- Existing ZIP
- The archive file itself must be under this size before compression starts.
- 1 GB
- Entries
- Files and folders combined. Larger archives are refused before compression.
- 100,000
What comes out
A standard ZIP, not a custom format.
The output uses DEFLATE compression and a conventional ZIP central directory. Windows Explorer, macOS Archive Utility, and common third-party archive tools can open it. There is no proprietary wrapper and no account-bound download.
Password-protected input is not supported. Encrypted archives are refused with a specific error rather than a generic failure. The current limits also keep typical output below the size where ZIP64 extensions become necessary.
Compression does not modify the bytes of a file. It changes how those bytes are represented inside the archive. When you open the ZIP, each original file is reconstructed and checked for integrity.
Cancel means cancel
A running job can be stopped. The page terminates the worker and discards the in-progress result. No partial archive is downloaded.
Now try it
See the size change for yourself.
Bring a folder or an existing ZIP. The result panel shows the real byte count before you download.