analyze-videos-with-frame-extraction-and-audio-context-in-claude

# Analyze videos with frame extraction and audio context in Claude Code Give Claude Code a video perception layer that extracts frames, transcribes audio, and lets Claude answer questions about local videos or YouTube URLs. ## Prerequisites Claude Code, Node.js 20+, ffmpeg, optional yt-dlp, Gemini API or Whisper/OpenAI audio backend ## Installation Use the upstream install or setup path that matches your environment: - git clone https://github.com/jordanrendric/claude-video-vision.git Requirements and caveats from upstream: - | **Local (Whisper)** | whisper.cpp or Python openai-whisper | Free, fully offline | brew install whisper-cpp + auto model download | - │ MCP Server (Node.js) │ - **Node.js 20+** (for the MCP server) Basic usage or getting-started notes: - ### 1. Install the plugin - Inside Claude Code, run these commands **one at a time**: - /plugin marketplace add https://github.com/jordanrendric/claude-video-vision - Source: https://github.com/jordanrendric/claude-video-vision - Extracted from upstream docs: https://raw.githubusercontent.com/jordanrendric/claude-video-vision/HEAD/README.md ## Documentation - https://github.com/jordanrendric/claude-video-vision ## Source - [Agent Skill Exchange](https://agentskillexchange.com/skills/analyze-videos-with-frame-extraction-and-audio-context-in-claude-code/)