VideoHighlighter

VideoHighlighter is a desktop app for finding useful moments in long recordings without uploading footage for analysis. It helps creators cut highlights, reviewers inspect camera recordings, and interview editors locate material worth keeping.

• Highlights with explanations: Score moments using scene changes, motion, audio, objects, actions and speech, then inspect a timeline and report that explain each selection and distinguish measured events from transcript claims.
• Visual search and event rules: Search footage with CLIP and define events through combinations of detections, such as a ball inside a net, object counts or conditions that persist over time.
• Transcripts and subtitles: Use local Whisper transcription to produce subtitles and chaptered videos, with optional translation that also runs locally.
• Reels and clips: Export a highlight reel and individual clips, or process GoPro and drone footage through a resumable workflow that copies recordings, selects highlights and matches music to the beat.
• Custom detection models: Teach a model a category by drawing a box in the player, train it on your footage, and share or obtain models through the model hub.

The app is free and open source under AGPL-3.0 and runs on Windows, macOS and Linux. It supports Intel, AMD and NVIDIA GPUs and uses Ollama for local language models. Analysis works offline, footage stays on your machine, and the basic pipeline doesn't require an API key. Ollama can also run on another computer on your local network.