A private image compressor you can verify

“No upload” is a claim, not a guarantee

A great many online compressors describe themselves as private. The words are on the page; the architecture behind them is not. Server-side compression is simply easier to build: the same code runs for everyone, batches are trivial, and file limits are easy to enforce. The cost is invisible to you — your photograph is copied onto someone else's machine, processed there, and hopefully deleted afterwards.

You cannot tell which kind you are using by looking at it. Both kinds show a drop zone, a quality slider and a download button. The difference is entirely in what happens underneath, which is why the promise is worth testing rather than trusting.

Verify it yourself in ten seconds

This method works on any image tool, including this one. It takes one photo and a browser you already have.

1. Open the page and press F12 (on macOS, Cmd + Option + I) to open developer tools.
2. Switch to the Network tab, tick Preserve log, then clear the list.
3. Drop in a photo, compress it, and download the result.
4. Read the request list that appeared.

What you will see: the page's own HTML, CSS and JavaScript, plus a WebAssembly codec fetched on demand when you pick a format that needs one. Those are pieces of the program — they are code, not your picture.

What you will not see: any request carrying your image. No upload, no POST, no multipart form. If a tool that advertises privacy shows a request containing your file at this step, then its privacy was marketing copy.

What actually happens to your file

Locally, the sequence is short: the browser decodes the image into raw pixels, optionally resamples it if you asked for a smaller size, re-encodes it in the format you chose, and hands you a download from memory. At no point does the data cross a network boundary, because there is nothing on the other end to cross it to.

One useful side effect: re-encoding copies pixels and nothing else, so EXIF metadata — including the GPS coordinates most phones write into every photo — is dropped. You do not have to remember to strip it; it simply does not survive the trip.

When this actually matters

• Identity documents. Passport scans, national ID cards, visa photos. These are the files you would least want sitting in a stranger's storage bucket.
• Medical images. X-rays, scans and clinical photographs carry some of the most sensitive data a person has.
• Work under an NDA. Unreleased product renders, client artwork, internal documents.
• Ordinary family photos. Faces, homes, school uniforms and embedded location data — plenty of people would rather those never left their laptop.

For a picture headed to a public website, a server-based compressor is perfectly adequate. The distinction only bites when the image is one that should not be seen by anyone else.

What this site does collect

Being precise matters more than being flattering, so here is the whole of it. There is one self-hosted, anonymous counter: your browser keeps a random identifier, hashes it, and sends that hash to our own /api/count endpoint at most once a day. It carries no filename, no image content and no IP-derived profile, and it exists only so the footer can show that anyone is here at all.

There is no Google Analytics and no third-party analytics script of any kind — nothing that would let an advertising network learn that you visited. If you would rather not be counted either, disconnect from the internet after the page loads: the compressor keeps working, because it was already downloaded.

If you need a specific target instead

Sometimes the requirement is not privacy but a hard number — a portal that rejects anything above 100 KB. The compress to 100 KB page does that with the same no-upload guarantee, and the main compressor switches between target size, fixed quality and full auto. iPhone photos in HEIC format go through HEIC to JPG.

Frequently asked questions