video-use

Why we included this project

video-use is for people who would rather describe an edit than click through a timeline. You drop raw takes in a folder, tell Claude Code or a similar agent what you want, and it returns a finished mp4 with filler words and dead space cut, color grading applied per segment, audio fades at every cut, and subtitles burned in. The agent works from a word-timestamped transcript plus a filmstrip-and-waveform image at decision points instead of raw frames, which keeps token costs reasonable. It also checks its own rendered output at each cut boundary and saves session state to a project file, so a later session can pick up where the last one stopped. For solo creators and small teams already comfortable with agent-driven workflows, that adds up to repeatable, conversation-driven post-production without presets or menus.