Why processing files in your browser is more private than uploading them

Last updated 2026-09-06

Search for "convert PDF to Word" or "compress image online" and the first page of results is full of sites that work the same way: you pick a file, it is uploaded to their server, processed there, and a download link appears. It is convenient, and for a holiday snap it hardly matters. For a CV with your address, a scan of your passport, a contract, a bank statement or a medical letter, it matters a great deal.

What happens on a typical converter site

1. Your file is transmitted in full to a server you know nothing about, in a jurisdiction you probably cannot identify.
2. It is stored on disk at least briefly, and often for "up to 24 hours" or longer according to the privacy policies of the largest services.
3. The converted result is stored too, and a download link is generated. Anyone with that link can fetch the file until it expires.
4. The service may log the file name, size, your IP address and the time. Some scan file contents for advertising or model-training purposes; their terms allow it.
5. Deletion, when it happens, is a promise you cannot audit.

None of this is necessarily malicious. It is just the consequence of the architecture. If a file leaves your device, its privacy depends entirely on someone else's practices, security and honesty.

The alternative: do the work on your own device

Modern browsers are capable of a great deal of computation. Three technologies make it possible to run file tools without a server:

• The Canvas API decodes images, scales them, draws on them and re-encodes them to JPG, PNG or WebP. Every image tool on this site is built on it.
• WebAssembly runs compiled code at near-native speed inside the browser sandbox. This is how a full FFmpeg video encoder, a PDF renderer and a DOCX parser can run on a web page without installing anything.
• Native compression streams give web pages access to the browser's own gzip engine, which is what powers the ZIP and GZIP tools here.

The web page is downloaded once, and from then on it works on files in your browser's memory. The result is handed to you through a normal download, generated locally. No file, and no part of a file, is transmitted. The site's server only ever sees a request for the page itself, exactly as it would for any static web page.

How to verify it yourself

You do not have to take a privacy policy's word for it. Two checks anyone can do:

1. Go offline. Load the tool page, then switch off wifi or enable aeroplane mode. Drop a file in and run the tool. If it works, nothing could have been uploaded. Every tool on this site passes this test, with one caveat: the video and audio tools download their FFmpeg engine on first use and need it cached before going offline.
2. Watch the network tab. In any desktop browser, open the developer tools (F12), select Network, and run a conversion. You will see requests for the page's own scripts and, if enabled, analytics and advertising beacons, but no request carrying your file. A server-side converter shows a large upload request the moment you drop a file.

The honest trade-offs

On-device processing is not magic. Because it uses your device's processor and memory:

• Very large jobs, such as a two-hour 4K video or a thousand-page scan, can be slow or run out of memory on a phone. A server with a powerful CPU would be faster for those.
• Some tasks need software that is simply too big or too proprietary to run in a browser. A perfect Word-to-PDF rendering, for example, really needs Word. The Word to PDF tool here is upfront about handling simple documents well and complex layouts imperfectly.
• The first use of the video and audio tools involves a one-time 32 MB engine download.

For the everyday jobs that make up almost all file conversions, compressing photos, converting between image formats, merging and splitting PDFs, zipping files, the browser is more than fast enough, and the privacy benefit is absolute rather than a matter of trust.

What this site does and does not collect

File contents: never, by design. The site uses anonymous analytics to see which tools are used, and is supported by advertising, both of which are described in the privacy policy and can be declined in the cookie settings. Neither has any access to the files you process, which never exist anywhere except your own browser.