This is a side project where I experimented with combining speech activity detection and scene changes to automatically extract short clips from long videos.
The goal was to avoid purely “viral score” models and instead generate clean, context-preserving clips that editors can tweak and find real highlights not just random cuts.
It’s free, no signup, and still very much a work in progress.
Would appreciate BRUTAL feedback from people working with video pipelines or ML-heavy media tools.
The goal was to avoid purely “viral score” models and instead generate clean, context-preserving clips that editors can tweak and find real highlights not just random cuts.
It’s free, no signup, and still very much a work in progress.
Would appreciate BRUTAL feedback from people working with video pipelines or ML-heavy media tools.