Last updated: August 2026
brainCUT is a browser-based video editor. Your footage is decoded, composited and encoded entirely on your device. We do not collect, store, or transmit your media or personal data to any server.
Media reaches the editor either through the browser's File System Access API, when you open a file or a folder, or through a file input and drag and drop in browsers without it. From there everything happens locally: frames are decoded with WebCodecs, composited with WebGL, and encoded back with WebCodecs when you export. No file data is uploaded to external servers.
Large files are left where they are and read a piece at a time. A copy is kept in the browser's own storage in two cases: when the file is 64 MB or smaller, so a project you reopen finds it without asking for permission again, and when the file arrived without a handle to it on disk, because then the copy is the only way back to it. A proxy or a converted copy, when you ask for one, is written to the same private area. All of these copies live on your machine and can be cleared with the site's data.
brainCUT uses your browser's own storage and nothing else. None of it leaves your browser, and all of it can be cleared at any time through your browser settings. This is everything the app writes:
localStorage holds small interface settings under five keys: braincut-preferences (everything in the preferences dialog), braincut-layout (panel sizes and which tabs are open), braincut-export (the export settings you last used), braincut-project-view (list or grid in the project panel) and braincut-effects-collapsed (which effect groups you have folded away).
IndexedDB holds two databases. braincut keeps your saved projects, the
recents list and which project was open last. braincut-media keeps three stores: sources, what points a project back at its footage — the file's name, size, last
modified date and the path it was picked from, plus a handle to it on disk, a copy of the file, or
both, following the rule above; peaks, the computed audio waveforms; and files,
used only in browsers that cannot write to the origin private file system directly.
The origin private file system holds the generated copies of your media: proxies (the smaller previews) and converted copies (files the browser could not decode as they were), plus scratch files written during an export. It is a private area of disk only this site can see.
Cache Storage holds the app itself, so it works offline: the files of the build you loaded, the last good copy of everything else, the ffmpeg core once it has been downloaded, and the one setting the service worker needs to know, which mirror the core may come from.
All of it stays inside your browser's storage for this site. Nothing here is sent anywhere, and clearing the site's data removes the lot, copies included.
We do not use tracking scripts, analytics services, or advertising networks.
Exactly two things ever talk to the network, and neither of them receives your media.
Fonts. The interface loads its fonts from Google Fonts (fonts.googleapis.com and fonts.gstatic.com) while the page loads. Google's privacy policy applies to that request.
The converter, only when you convert. Some files use codecs the browser cannot decode on its own, HEVC and ProRes among them. Converting one of those downloads the ffmpeg.wasm core from jsDelivr, once, and keeps it in the browser cache for later. Only the core's own file names are requested, never your media. If you never convert a file, this request never happens. The mirror can be changed or emptied entirely through the ffmpegMirror preference, which turns conversion off.
That is the complete list. There is no backend, no API, no error reporting and no update check. Once the page has loaded, the editor works with the network switched off.
brainCUT does not use cookies, and there is no account to sign in to. The interface settings listed above are stored in localStorage.
We may update this policy from time to time. Changes will be reflected on this page with an updated date.
For questions about this policy, contact us via GitHub.