# video-use

A tool that edits raw video footage into a finished cut through a conversation with a coding agent such as Claude Code, handling filler-word removal, color grading, subtitles and animated overlays.

video-use edits video by chatting with a coding agent: drop raw footage into a folder, describe the edit to Claude Code or a similar agent, and get a finished file back. It works on any kind of footage, such as talking heads, montages, tutorials, travel clips or interviews, without presets or menus.

It cuts filler words and dead space between takes, auto color-grades each segment, adds short audio fades at every cut, burns in customizable subtitles, and can generate animation overlays through tools such as HyperFrames, Remotion, Manim or PIL, run as parallel sub-agents. It self-checks the rendered output at each cut boundary before showing it to the user, and keeps session memory in a project file so a later session can continue where the previous one left off.

It reads video through a word-level audio transcript with speaker diarization and audio-event tags, generated by one call per source to ElevenLabs Scribe, plus an on-demand visual composite of filmstrip, waveform and word timing. It is written in Python, requires an ElevenLabs API key, and is open source.

Website: https://github.com/browser-use/video-use