Your video and audio stay local
Importing a local video makes it available to the editor running in your browser. The current app does not upload that original file for editing, automatic speech recognition, face tracking or final video rendering.
Whisper speech model files download from Hugging Face through a Cloudflare cache. MediaPipe face-model and runtime files download from Google and jsDelivr. These services receive normal network request information, such as your IP address when your browser connects to them; local video content is not included in those model requests.
Optional features make cloud requests
When you choose AI highlights in Find highlights, the app sends transcript text, source timings and your topic preference to our Cloudflare Worker and Workers AI. Choose Quick suggestions if you do not want that transcript request.
YouTube preview sends the video URL to the Worker, which requests metadata from YouTube. It retrieves a title and thumbnail, not the video media. Loading the thumbnail also makes a request to its hosting service.
The website runs on Cloudflare. Hosting requests include normal connection and request information. Worker observability is enabled, and failed API requests may generate operational error logs. A request limiter uses IP addresses; a separate database stores a shared daily AI request count. This preview has no advertising tracker or analytics script added by ClippingShorts.
Project recovery is your responsibility
The latest project is saved in this browser’s local storage. Downloaded project JSON files include the transcript, clip ranges and editing settings. They do not contain the original video. Keep both the original file and a saved project if you want to continue later.
You can remove the browser’s stored project by clearing this site’s storage in browser settings. Downloaded files remain on your device until you remove them. There is no account-based cloud media library, collaboration workspace or server-side export archive in this preview.
What is available in this preview
- Local import, editable clips, captions, manual reframing, largest-face tracking and MP4/WebM export.
- Optional cloud AI highlights, with 25 requests per UTC day shared across the public preview.
- Batch ZIP exports limited to ten clips and ten minutes total to reduce memory pressure.
- YouTube metadata preview only. Direct downloading, billing, team accounts and social publishing are not connected.
Browser codecs and available memory affect which files can be decoded and exported. Speech recognition and AI highlight suggestions can make mistakes. Review your captions, cropping and source context before publishing a result.