Privacy
Short, because there is not much to say.
Your files
They are never uploaded. Every tool runs FFmpeg inside your browser, so your media never leaves the device. This is not a policy choice that could be changed later without you noticing: there is no server-side component that could receive a file. See how it works, and verify it in your browser's network panel.
The tools that use a model
Captions are transcribed by a Whisper model, and the two image checkers read pictures with a vision model. All of them run inside your browser, so a recording or a photograph is no more uploaded than a video is. One caveat is worth stating plainly: the first time you use one of those tools, its weights are downloaded from huggingface.co. That request tells them your IP address and which model you asked for, the same as loading an image from any website. It carries none of your audio, video, text or pictures, and it happens once per model because the browser caches the result. If you would rather it did not happen at all, the site can be configured to serve the weights from its own domain.
What the site stores on your device
- The FFmpeg WebAssembly file, in the normal browser cache, so it does not download again.
- The Whisper speech model, cached the same way, once you have used a caption tool.
- An image model, cached the same way, once you have used one of the image checkers.
- Nothing else. No cookies are set by this site, and no login exists.
Server logs
The site is static files served by a web server, which keeps standard access logs: IP address, timestamp, the page requested, and user agent. They are used for nothing but keeping the site up and are not sold or shared.
Third parties
No analytics, no ads, no fonts loaded from anyone else's server, and no embedded scripts of any kind. The one exception is the model download described above.