Clipify identifies compelling moments within video files, ranging from humorous dialogue in podcasts to repeated impact actions like drumbeats or hits. It streamlines the process of extracting standalone clips for social media by automating the technical overhead of reframing and captioning.
Content creators and producers can reformat 16:9 footage to a 9:16 vertical aspect ratio, apply time-warp pans for tighter reveals, and burn word-by-word captions directly into the output.
Follow the documented setup, then try a first task.
Installs: The single Clipify skill and its bundled scripts · Claude Code
Platform: macOS; Linux and Windows require removing the documented -hwaccel videotoolbox flags
Run from: /tmp/clipify/
In a terminal, install ffmpeg with libx264 using the author-documented command.
brew install ffmpeg
In a terminal, install Whisper using the author-documented command.
pip install openai-whisper
After ensuring Python 3 is installed, run the author-documented command in a terminal to install numpy.
pip install numpy
In a terminal, clone Clipify into Claude Code’s skills directory.
git clone https://github.com/louisedesadeleer/clipify.git ~/.claude/skills/clipify
Restart Claude Code. When using Clipify, create and use /tmp/clipify/ as its working directory and leave its artifacts there for debugging.
Uses clipify
Use the clipify skill in its default dialogue mode with the video at <video-file-path>; replace <video-file-path> with the actual local video path. Transcribe it and propose 3–5 clip-worthy moments with timestamps and titles. Stop after presenting the candidates without cutting or rendering anything yet.
Claude