Pick several files and get back one compressed archive, built in this tab. Text and documents shrink substantially. Photos and PDFs will not, and this page is straightforward about which is which.
Runs in your browser. Nothing is uploaded.
Building an archive here reads each file you picked, runs its bytes through DEFLATE at level 6, and writes the compressed results into one container with a directory at the end recording what went where. DEFLATE works by finding repeated sequences and replacing later copies with a short reference back to the first. That is why it does so well on text: prose, source code, CSV exports and log files repeat words, indentation and punctuation constantly. Three plain text files totalling 167,472 bytes came out as a 27,859 byte archive, a saving of 83.4 percent, in 17 milliseconds.
It is also why it does nothing for a folder of photographs. A JPEG, a PNG, a PDF containing images, an MP4 and an existing ZIP have all already had their redundancy removed by a codec designed for exactly that data, and there is no repetition left for a general-purpose algorithm to find. A JPEG and two PDFs totalling 2,273,612 bytes produced an archive of 2,269,170 bytes: a 0.2 percent saving for 101 milliseconds of effort. Zipping those files is still worth doing, because one attachment is easier to send than three, but expecting it to make them smaller is the most common misunderstanding about archives.
Choose files to zip
The same questions a security review would ask, answered plainly.
The parts that matter when the files are sensitive or numerous.
Select everything in one go and get a single archive back. This is the actual reason most people zip things: mail clients and upload forms deal far better with one file than with fifteen.
DEFLATE at level 6, the same setting zlib uses by default. Text-shaped data typically loses four fifths of its size, which is the difference between an archive being useful and being a container.
Two files called report.pdf from different folders both go in, with the second renamed report (1).pdf rather than silently overwriting the first.
The archive is assembled in the page from files read off your disk. For anything confidential this is the difference between a two-minute task and a conversation with whoever owns the data policy.
There is no upload step and no server to receive one. The page loads once, then all work on your archives happens in local JavaScript.
No daily limits, no per-file ceilings imposed by a server tier, no watermarks, and no account. The practical limit is your own device's memory.
The same settings against very different data, and why the results diverge so sharply.
The case where compression earns its keep.
Input
3 plain text files, 167,472 bytes total
archive.zip
27,859 bytes, 83.4% smaller, 17 ms
Settings: DEFLATE, level 6
Text repeats itself constantly, so most of the file can be replaced with references to earlier occurrences. Source code, CSV exports and log files behave the same way and see similar savings.
The case people expect to work and it does not.
Input
1 JPEG and 2 PDFs, 2,273,612 bytes total
archive.zip
2,269,170 bytes, 0.2% smaller, 101 ms
Settings: DEFLATE, level 6
These formats are already compressed by codecs built for their own data, so there is nothing left for DEFLATE to remove. The archive is still one file instead of three, which is usually the real goal.
Real files, real numbers, and the machine they were run on.
| Scenario | Input | Output | Time | Notes |
|---|---|---|---|---|
| three plain text files | 167,472 bytes | 27,859 bytes | 17 ms | 83.4% smaller |
| a JPEG and two PDFs, all already compressed | 2,273,612 bytes | 2,269,170 bytes | 101 ms | 0.2% smaller, and that is expected |
How this was measured: Measured by driving this page in headless Chrome 152 on an Apple M4 Pro running macOS 15.5, on 2026-09-17. The fixture is handed to the page's own file input, the page's own button is pressed, and the figure recorded is the size of the file the page produces. Timings come from a warm page on a fast machine with no network involved, so treat them as a floor rather than a typical result.
Add them all and create the archive. Whether it gets under an attachment limit depends entirely on what the files are: a folder of documents usually will, a folder of photographs almost certainly will not, because photos are already compressed.
This page does it without anything installed. ChromeOS can create archives from the Files app, but on a managed device that is sometimes restricted, and a browser tab is not.
That is not offered here, and for most inputs it does not matter: already-compressed files come out essentially unchanged anyway. Compression is applied uniformly because the cost on incompressible data is milliseconds.
Select the files inside it. The picker takes files rather than directories, so a deep folder tree is better handled by your operating system's own right-click option, which preserves the structure.
One attachment instead of twelve, and for text-shaped documents a much smaller one. The recipient gets a single file that every operating system can open without extra software.
Contracts, HR files and financial records get bundled before they are sent. Building the archive locally means the sensitive part of the job never involves a third party at all.
Log files and CSV dumps are close to the ideal case for DEFLATE, routinely losing most of their size. Zipping them before storage is the cheapest space saving available.
Portals that accept a single file force everything into one archive. Doing that in a browser tab works on a university machine where installing software is not an option.
Source files compress extremely well and stay exactly as they were, so a zipped snapshot is a reliable way to hand over code without a repository.
The page keeps working with no connection once it has loaded, so an archive can be prepared on a train and sent when there is signal again.
Including the one that surprises people most often.
Files are added by name at the top level of the archive, so a selection spanning several directories comes out flat. Your operating system's own compress option is the better choice when the tree matters.
Archives are written unencrypted. Anything requiring a password needs a desktop tool, and the extractor here cannot open encrypted archives either, so this is consistent rather than an oversight.
Each file is read into the tab and the finished archive is assembled there too, so the working set is roughly the total size twice over. Large jobs behave much better on a laptop than a phone.
Photos, PDFs, video and existing archives come out about the same size, because their redundancy has already been removed. Compress the images first if the goal is a smaller total.
No. The files are read from your disk by the page and compressed in the tab, and the finished archive is handed straight to your downloads. Open your browser's Network panel while you do it and you will see no requests at all after the page itself has loaded.
Because JPEG and PNG have already done the compressing. DEFLATE looks for repeated byte sequences, and an image codec has already removed those. A JPEG and two PDFs went from 2,273,612 bytes to 2,269,170, a saving of 0.2 percent. The archive is still useful as one file rather than three.
It depends entirely on the type. Plain text, CSV, JSON, HTML, source code and logs typically lose 70 to 85 percent; measured here, three text files went from 167,472 to 27,859 bytes. Office documents are already zipped internally and save little. Images, audio, video and PDFs save almost nothing.
DEFLATE at level 6, which is zlib's default and the point where extra effort stops buying meaningful size. There is no setting to change it, because on compressible data the gain from level 9 is small and on incompressible data no level helps.
No. Archives are created without encryption and there is no option to add one. If the contents need protecting in transit, encrypt them with a desktop archiver or send them through a channel that is already encrypted.
Both are kept. The second is renamed with a numeric suffix, so two files called report.pdf become report.pdf and report (1).pdf. Nothing is silently replaced, which matters when you have selected the same filename from two different folders.
Nothing is capped by this page, because there is no server applying a quota. The real limit is memory: the files and the archive being built both sit in the tab at once. A few hundred megabytes is reasonable on a desktop machine.
Yes. It is a standard ZIP using DEFLATE, which every operating system has opened natively for years, along with every archive utility. Nothing unusual is written into the container.
The file picker takes files, not directories, so you select the contents rather than the folder. Note that the paths are not rebuilt inside the archive, so everything arrives at the top level.
No. DEFLATE is lossless, so extracting the archive gives back byte-for-byte what went in. This is unlike image or video compression, where making something smaller means discarding detail.
Because the algorithm still has to read every byte looking for repetition, even when it finds almost none. The JPEG and PDF job took 101 milliseconds against 17 for the much smaller text files, and produced almost no saving for that work.
Yes, once loaded. Compression happens entirely in your browser, so after the first visit you can disconnect and still build archives. A fresh reload while offline needs the page to still be in the browser cache.
There is no limit imposed by us, because there is no server tier to enforce one. The real constraint is your device's available memory, since the whole file is held in RAM while it is processed. Very large files are slower on phones than on a laptop.
No. There is no sign-up, no email, and no usage tracking tied to an identity. The tool works the first time you open it.