Why "never uploaded" is the whole design
Search for any of these conversions and you will find dozens of sites that do it. Almost all of them upload your file to a server, convert it there, and hand it back. That means your photo — with its location data, its faces, whatever is in it — sits on a machine belonging to someone you have never met, for a period of time they decide. Most of the time nothing bad happens. But you have no way to verify that, and for a lot of people that is reason enough to look elsewhere.
These tools take the other approach. Your browser already contains decoders for JPEG, PNG, WebP and AVIF, and a WebAssembly build of libheif or libavif closes the gap for HEIC and older browsers. Encoding happens on a <canvas> in the tab. The consequence is that there is no server in the loop at all — which is why every page here ships a panel that counts the requests it makes and adds up the size of every outbound payload, so you can check the claim instead of trusting it.
Nothing to store or leak
There is no upload, so there is no queue, no retention policy and no third party holding your images.
No queue, no limits
Your own CPU does the work, so there is no file-size cap, no hourly allowance and no waiting behind other people.
Free with no catch
No account, no email capture, no watermark, no upgrade nag. There is no per-conversion cost to recover because there is no server cost.