Video Understand

This handles video transcription and understanding across nine different providers with automatic fallback. You throw a YouTube URL or local file at it, and it picks the best available option based on your API keys, preferring full video analysis (Gemini, Vertex AI, OpenRouter) over audio-only transcription. The smart part is the fallback chain: if you don't have API keys configured, it drops down to FFMPEG with scene detection or local Whisper. Each provider supports multiple models, output is normalized JSON across all of them, and it handles the annoying bits like macOS path issues and format conversion. Useful when you need to extract information from video content without manually configuring which service to use.

skills.sh installs: 305
GitHub stars: 34
last updated: Jun 3, 2026

View on GitHub

Install

$ npx -y skills add jrusso1020/video-understand-skills --skill video-understand --agent claude-code

Installs into .claude/skills of the current project.

Files

Video Understanding

Multi-provider video understanding with automatic fallback and model selection.

Quick Start

# Check available providers
python3 scripts/check_providers.py

# Process a video (auto-selects best provider)
python3 scripts/process_video.py "https://youtube.com/watch?v=..."
python3 scripts/process_video.py /path/to/video.mp4

# Custom prompt
python3 scripts/process_video.py video.mp4 -p "List all products shown with timestamps"

# Use specific provider/model
python3 scripts/process_video.py video.mp4 --provider openrouter -m google/gemini-3-pro-preview

# List available models
python3 scripts/process_video.py --list-models