How to compress images without uploading them

What "processed locally" actually means, why it matters more for some photos than others, and how to verify it yourself.

The short answer

Use a tool that runs entirely in your browser instead of one that sends your file to a server. Browsers have built-in APIs (createImageBitmap() to decode an image and canvas.toBlob() to re-encode it) that are capable enough to compress, resize, and convert most images without any network request carrying the file. IMAGE SIZE PRO's compressor works this way; you can confirm it yourself with the steps below.

Why "upload, then we delete it" isn't the same thing

A lot of free image tools are upfront about deleting your file after processing, often within hours. That's a reasonable-sounding promise, but it still requires a real upload to happen first, meaning your file briefly exists on infrastructure you don't control, is subject to whatever that company's actual practices are (not just their stated policy), and could be logged, cached, or intercepted in transit before deletion ever happens. A tool that never uploads the file in the first place removes that entire window, rather than shortening it.

How to check whether a tool actually processes locally

Open your browser's developer tools (usually F12, or right-click → Inspect) and switch to the Network tab.
Clear the log, then add an image to the tool and trigger the compress/resize action.
Watch what appears. A local tool shows no new request containing your image's data. An upload-based tool shows a request (often a POST to an API or file-upload endpoint) with your file in the payload, right at the moment you click compress.

This takes under a minute and doesn't require any technical background beyond knowing what to look for. It's worth doing once for any tool you're about to hand a sensitive photo to.

When this is worth caring about

For a casual photo you're posting publicly anyway, it doesn't matter much which kind of tool you use. It matters considerably more for passport photos, visa documents, ID scans, medical forms, or anything else tied to your identity. That's exactly the category of image people most often need to force under a strict file-size limit for a government or institutional portal, and exactly the category you'd least want sitting on a third-party server even briefly.

Common mistakes

Trusting a privacy claim without checking it. "Your privacy matters to us" in a footer is marketing copy, not a technical guarantee. The network-tab check above takes the guesswork out.
Assuming any free tool must be upload-based. It's the opposite, in one sense: local processing is usually cheaper to run than a server-side pipeline, since the site doesn't pay for compute or storage per image.
Confusing "no upload" with "works offline." The page itself still has to load once over the network. Local processing means the image data doesn't travel, not that an internet connection is never needed.

Common questions

Is a browser-based compressor as capable as an upload-based one?

For the core job (compressing, resizing, and converting between JPG, PNG, and WebP), yes, because modern browsers ship the same image codecs a server would use. Where upload-based tools sometimes pull ahead is in more exotic operations, like AI-based background removal or server-side batch pipelines processing thousands of files, which do need real compute beyond what runs in a tab.

How can I tell if a tool is actually processing locally, or just claiming to?

Open your browser's developer tools, go to the Network tab, and process an image. If the tool is genuinely local, you'll see no outgoing request carrying your image data, just the page's own assets loading once at the start. If it's upload-based, you'll see a request (often to an API or CDN endpoint) with your file in the payload right when you click compress.

Does local processing mean it works without an internet connection?

Not automatically. The page itself still needs to load once over the network. After that, the actual compress/resize/convert step doesn't require a connection, but whether the tool works fully offline afterward depends on that specific site's setup, not on local processing alone.