Search for a way to shrink a passport photo and you will find dozens of free tools. Nearly all of them work the same way: you upload your photo, a server processes it, and you download the result.
For a holiday snap, fine. For a passport photo, a signature, a bank statement or a scan of your ID card, it is worth stopping to ask what you just did, because the answer is that you sent a government identity document to a machine you know nothing about, run by people you cannot name, in a country you did not choose.
What actually happens when you upload
Your file travels to a server. There, at minimum, it is:
- Written to disk, because that is how web servers handle uploads
- Processed by software you cannot inspect
- Written to disk again as the result
- Served back to you over a link that may remain valid for some time
Along the way it may pass through a CDN, be logged by a reverse proxy, be included in a routine backup, or sit in a temporary directory that a cleanup job clears, eventually. Most sites promise deletion after an hour or a day. Some mean it. You have no way to check either way.
None of this requires bad intent to go wrong. A misconfigured storage bucket, an unpatched server, a casually-handled backup, the file only has to be exposed once.
Why free tools are free
Running servers costs money. Bandwidth for millions of image uploads costs a lot of money. A free tool is being paid for somehow, and it is worth knowing which way.
Advertising is the honest answer, and it is the most common one, including here. The site shows ads, the ads pay for hosting, everyone understands the deal.
The less honest answers involve the files themselves: retained for training data, analysed for whatever can be extracted, or simply kept because storage is cheap and nobody wrote a deletion job. With identity documents the downside is not abstract. A passport scan is enough to open accounts in your name.
You cannot tell which kind of site you are on by looking at it. A privacy policy is a statement of intent, not evidence.
There is a structural fix
Modern browsers can do all of this work themselves. The canvas API can decode, crop, resize and re-encode images. JavaScript can read and rebuild PDFs. None of it needs a server.
When a tool runs entirely in the browser, the question "what does the server do with my file?" disappears, because there is no server and no file transfer. Your photo is read by code running on your own machine and never leaves it.
This is not a promise you have to take on trust. It is a property of how the page works, and you can verify it yourself.
How to tell the difference
Three checks, in increasing order of effort:
1. Disconnect from the internet
The simplest and most convincing test. Load the page, turn off your Wi-Fi, then use the tool. If it still works, nothing is being uploaded, it cannot be. If it fails or hangs, your file was going somewhere.
Try it on this site. The file tools keep working offline once loaded, because all the processing code is already in your browser. (The AI tools download their model the first time you use them.)
2. Watch the network tab
Open your browser's developer tools, go to the Network tab, and use the tool. An upload appears as a large POST request with your file as the payload. If the only requests are for page assets, nothing left your machine.
3. Look for the tell-tale signs of a server
Server-based tools tend to have characteristics that in-browser tools do not need:
- A file size limit, because upload bandwidth costs money. In-browser tools have no reason to impose one.
- A queue or progress bar for the upload itself, as distinct from processing.
- A results link rather than an immediate download.
- Sign-up or daily usage limits, rationing a server resource.
What in-browser processing cannot do
Being honest about the limits, since a tool that oversells itself is its own warning sign:
- HEIC. Only Safari opens Apple's HEIC format by itself, so many sites convert it on their server, which means they receive your photo. The Daily Toolbox's HEIC to JPG reads it in your browser instead, or set your iPhone to Settings → Camera → Formats → Most Compatible.
- OCR and text recognition. Possible in a browser but heavy. Most sites offering it use a server.
- Very large documents. Processing happens in your device's memory, so a 500-page scan will struggle on a phone. A server has more memory, that is a real trade-off, not a trick.
Sensible habits regardless
- Send only the pages needed. A bank statement proving your address also lists every transaction. Extract the single page.
- Crop tightly. Backgrounds contain more than you notice, documents on a desk, a view through a window.
- Strip metadata. Phone photos carry GPS coordinates. A photo taken at home has your home address in it. Any tool that redraws the image, including all of ours, removes this.
- Watermark copies where appropriate. If you must send an ID scan to a third party, marking it with the purpose and date limits its reuse.
- Prefer the official portal. If a government site has its own photo tool, use it, it is at least accountable.
The point
This is not an argument that every online tool is malicious. Most are ordinary businesses doing ordinary things. It is an argument that for a small category of genuinely sensitive files, the safest architecture is the one where the question never arises.
Everything on this site works that way, and you can prove it in ten seconds by turning off your Wi-Fi. That is a better assurance than any privacy policy, including ours.