The default settings target the dead air common in talking-head videos and screen recordings. Raise the silence threshold when steady room noise hides pauses, or lower it when quiet speech is being mistaken for silence. Each detected range stays selectable, so you can preserve a deliberate pause instead of accepting an all-or-nothing edit.
Analysis and export use a WebAssembly build of FFmpeg in this tab. Export re-encodes the kept audio and video sections together as H.264 video with AAC audio, which keeps jump cuts synchronized and produces a broadly playable MP4.