Every tool on Filefel is marked either on-device or server before you pick a file. 15 of the 15tools live today are on-device: the file you choose is read into your browser tab and processed there with WebAssembly and the browser's own codecs. It is never uploaded, never sent to us, and never touches a server we run. You can confirm this yourself by opening your browser's network panel while a job runs.
Nothing about your files persists anywhere. There is no IndexedDB cache, no service-worker copy, and no server-side log of a filename or file content. Results live in memory and disappear the moment you close the tab — the product tells you that on the page itself, not just here.
Analytics. Filefel uses Google Analytics 4 to see which pages and tools get used — page views, and whether a tool was started, completed or failed. Nothing about your files — not a filename, not their contents — is ever part of that data.
No accounts.There is no sign-up, no email collection, and no usage cap tied to an identity. Nothing to leak because there's nothing to store.
Some tools we add later will genuinely need a server — anything that calls a model, for example. Those will say so on the page, before you choose a file, using a visually distinct badge so it's never confused with the on-device promise.
Questions about this page can go to the contact page.