Privacy
Your files never leave your device
ClipLime was built around one rule: we never receive the videos, recordings, images or text you work with. Here's how that works, in plain language.
What happens when you open a file
When you choose or drop a file, your browser gives this page permission to read it from your disk. The tool then decodes, edits and re-encodes it using engines that are built into your browser (the WebCodecs and Web Audio APIs), on your own processor and graphics card. The finished file is handed back to your browser's download manager. At no point is any part of your file sent over the network.
You can check this yourself: open your browser's developer tools, switch to the Network tab and run any tool. You'll see page assets load, but you won't see your file go anywhere.
The AI subtitle model
The subtitle generator needs a speech-recognition model. The first time you use it, your browser downloads an open-source Whisper model from the Hugging Face content delivery network and caches it. This is a one-way download of the model; your audio is transcribed locally and is not sent to Hugging Face or to us.
The face detector
When the video repurposer follows faces, your browser downloads a small open-source face detector the first time: the MediaPipe code from the jsDelivr content delivery network and a model file from Google Cloud Storage. Your browser caches both. The video frames are checked on your device and are not sent to jsDelivr, Google or us. If you turn on captions there, the same Whisper model as above is used.
YouTube thumbnails
The thumbnail downloader loads publicly available thumbnail images directly from YouTube's image servers in your browser. We don't proxy or log those requests.
The footer badge
The footer shows a “Featured on Launch Llama Tools” badge. When you scroll to it, your browser loads that image from tools.launchllama.co like any image on the web, so their server sees a standard request for it. No file data or tool activity is part of that request.
The files-processed counter
The counter in the top bar, on the home page and in the footer shows how many files ClipLime has finished. When a tool finishes a job, your browser sends our server one word for the kind of file (video, audio, image, pdf or data) and, when a batch finishes several files at once, how many. We add that to a running total. Only a finished file or job counts: opening a page or a tool, adding a file or changing a setting never does, and calculators, text tools and QR codes are not counted at all. We store only the totals per kind and a count per day. We don't receive or keep file names, sizes, contents, what you did to the file, or anything that identifies you. To stop someone inflating the number, the server briefly remembers in memory how many counts came from one address in the last minute, and forgets it right after.
What we do collect
The site is served as static pages. Like any website, the hosting provider may keep standard, short-lived server logs (such as IP address, browser type and the page requested) to keep the service secure and running. We don't use advertising trackers, we don't sell data, and we don't build profiles.
Some tools remember small preferences, such as your teleprompter speed, in your browser's local storage so they're there next time. That data stays on your device and you can clear it at any time from your browser settings.
Camera, microphone and screen
The recorders ask for camera, microphone or screen access only when you press record, and only for the recording you start. Recordings are kept in memory in your tab until you save them, and are discarded when you close the page.
Contact
Questions about privacy? Email [email protected].