31 August 2026
Why your files never leave your device
Every tool on Khallas runs inside your browser, on your own machine. That is a statement about how the code is built rather than a promise in a policy, and you can check it in about a minute. It also carries real limits worth knowing before you rely on it.
What actually happens when you drop a file
When you pick a file in a browser, the browser grants the page permission to read it from your disk. It does not send it anywhere. The code that processes it downloaded with the page before you chose anything: JavaScript and WebAssembly, running inside the tab on your own processor and memory. The pages are read, the merge or compression or redaction you asked for is performed, a new file is written into browser memory, and you download it. At no point in that chain does your document cross the network. Many popular tools work the opposite way, uploading the file to a server, processing it there and returning a download link.
Never uploaded also means never crossing a border
Most privacy questions people ask about web tools are really questions about a server: where is my file stored, how long does it stay, who can reach it, does it leave the country. When the file never leaves your device, none of those questions has anything to attach to. There is no server receiving it, no copy kept for an hour or a day before deletion, no transfer between countries, and no deletion promise you have to trust. This describes how the tool is built rather than offering legal advice, but it takes the whole subject off the table instead of answering it.
The documents where this actually matters
Uploading a public brochure to somebody's server changes nothing. The difference shows up with a payslip, a bank statement, a contract, an ID copy, a medical report or an employee file. Our redaction tool removes the text under a marked area for real, by re-rendering the page as an image rather than covering it over. Our protect tool adds a password using AES-128, a standard part of the PDF specification. Both work the same way as everything else here: on your device. The sensitive document is never on its way to anyone at any moment, before processing or after it.
How to check instead of taking our word for it
Open your browser's developer tools, choose the network tab, run any tool on a real file and watch what leaves. You will see requests for the code and fonts the tool needs, and no request carrying your document. OCR is the one visible exception: on first use the recognition engine and language model download, a few megabytes in total, and stay cached in your browser for later runs. We serve them from our own servers rather than a third-party host, so no request reaches anyone else while your document is being read. The test takes a minute and saves you trusting any marketing sentence, ours included.
The honest limits: your device does the work
The other side of the coin is that your hardware is the ceiling. Very large files need memory, and an older phone may crawl or give up where a recent laptop sails through. OCR is the heaviest tool here and reads up to twenty pages per run, and batch processing takes up to three files at a time. A tool that uploads to a powerful server will handle enormous files more steadily, and that is a genuine trade-off rather than something we hide: you give up some capacity in exchange for the file staying put. If capacity matters more than privacy on a particular job, the other option suits you better, and you deserve to learn that here.
The one exception we have planned
We may one day add a single feature that needs server-side processing because a browser genuinely cannot do it. If that happens, the terms are settled in advance: entirely optional, with explicit consent for each individual file every time, stated where you are standing rather than buried in a policy, and never the default. Everything else stays inside your browser exactly as it works today. We are writing this now rather than at launch, so that never uploaded cannot quietly change meaning later. No such feature exists today, and not one tool on the site sends your file anywhere.
Frequently asked questions
How can I verify my file was not uploaded?
Do not take the promise, test it. Open your browser's developer tools, go to the network tab, and run a tool on a real file. You will see requests for the page's own code and fonts, and none carrying your document. The one visible exception is the OCR engine downloading on first use, which is a program coming down to you rather than a file going up from you.
Does this mean my file is completely safe?
It means we are not part of the risk at all, because the file never reached us. The risks on your own device remain: who else uses it, what is installed on it, and where you save the output. And the file does leave your device the moment you email it or send it on WhatsApp, which is entirely outside anything we control.
