Why your files should never leave your device to be converted
The classic online converter works like this. You pick a file, it travels across the internet to a server, the server runs the conversion, and the result travels back. That design made sense when browsers could barely resize an image. It has not been necessary for several years, and it carries three costs that are easy to overlook.
You wait twice
Every conversion includes a full upload and a full download. On a typical home connection with 20 Mbit upload, a 500 MB video takes over three minutes to upload before any work begins. The conversion itself might take twenty seconds. Most of the time you spend waiting is transfer, not processing.
Your file sits on someone else's machine
Whatever the privacy policy says, the file was on hardware you do not control, handled by software you cannot inspect, for some period of time. For a meme that is irrelevant. For a signed contract, a medical scan, an unreleased product video or a passport photograph, it is a real consideration, and most people do not stop to make that distinction before dragging a file in.
Limits exist to protect the server, not you
The 100 MB cap, the daily conversion limit and the queue during busy hours all exist because processing files for thousands of strangers is expensive. That cost is then recovered through a subscription. The limits are a business constraint dressed up as a product feature.
What changed
WebAssembly lets browsers run compiled code at close to native speed. FFmpeg, the tool that underpins most of the video and audio world, has been compiled to WebAssembly and runs inside a browser tab. The same is true for image and PDF processing libraries. The browser is now a capable media workstation.
Once the work happens locally, the three costs disappear at once. There is no upload, so conversion starts immediately. The file never leaves the machine, so privacy is not a policy question. There is no server bill, so there is no reason to impose limits or ask for a subscription.
The honest trade-offs
- The engine has to be downloaded the first time, roughly 30 MB. It is then cached, so only the first conversion of the session waits.
- Very large files are limited by your device memory rather than by a server. On a modern laptop that means around 1 to 2 GB.
- A ten year old phone will be slower than a data centre. For most files the difference is seconds.
Those are acceptable prices for never uploading a file again. GIF Diet is built entirely on this model, which is also why it has no accounts, no limits and no paid tier.