What problem does it solve?
This Skill provides a comprehensive toolkit for analyzing and editing video content by using FFmpeg for frame extraction, Whisper for local speech transcription, Gemini Audio for deep audio understanding, and optional Shazam for track identification. It enables automated analysis workflows and prepares structured outputs.
Core Features & Use Cases
- Video analysis and editing: clip, merge, and split videos while generating visual frames and synchronized transcripts.
- Multi-modal understanding: extract frames at intervals or scene changes, transcribe speech, detect non-speech audio (music, ambient sounds), and summarize video content.
- Music identification integration: detect music and optionally identify tracks with Shazam to attach titles, artists, and albums.
- Use Case: Given a video with a product demo, the skill can extract key frames, generate a transcript, identify any background music, and prepare a summary for review or publication.
Quick Start
Run the video-toolkit on a video file to automatically extract frames, transcribe audio, analyze music presence, and produce a summary. For example: bash plugins/video-toolkit/skills/video-toolkit/scripts/analyze_video.sh /path/to/video.mp4
