video-analyzer-mcp 0.1.0

MCP server for video analysis - extracts frames, transcribes audio, analyzes visuals with Claude, and generates storyboard breakdowns

Video Analyzer MCP Server

An MCP (Model Context Protocol) server that analyzes videos and generates storyboard breakdowns. Extract frames, transcribe audio, analyze visuals with Claude Vision, and compute stylistic fingerprints — all usable directly from Claude Desktop or Claude Code.

Tools

Tool	Description
video_analyze	Full pipeline: download, extract frames, transcribe, visual analysis, stylistic fingerprint, storyboard document
video_extract_frames	Extract representative frames using scene detection or fixed intervals
video_transcribe	Transcribe audio with OpenAI Whisper (timestamped, speaker-labeled)
video_fingerprint	Generate an 8-field Stylistic Fingerprint v3 classification
video_check_deps	Verify all required dependencies are installed

Prerequisites

• Python 3.10+
• FFmpeg — brew install ffmpeg (macOS) or apt-get install ffmpeg (Linux)
• ANTHROPIC_API_KEY — set as an environment variable

Installation

From PyPI

pip install video-analyzer-mcp

For Claude Desktop

Add to your Claude Desktop config (~/Library/Application Support/Claude/claude_desktop_config.json):

{
  "mcpServers": {
    "video-analyzer": {
      "command": "uvx",
      "args": ["video-analyzer-mcp"],
      "env": {
        "ANTHROPIC_API_KEY": "your-key-here"
      }
    }
  }
}

For Claude Code

claude mcp add video-analyzer -s user -- uvx video-analyzer-mcp

Usage Examples

Once installed, you can ask Claude:

• "Analyze this video and create a storyboard" — runs video_analyze
• "Extract frames from this YouTube video" — runs video_extract_frames
• "Transcribe the audio from this video" — runs video_transcribe
• "What's the stylistic fingerprint of this video?" — runs video_fingerprint
• "Check if video analyzer dependencies are installed" — runs video_check_deps

Stylistic Fingerprint Fields

The fingerprint classifier produces 8 deterministic fields:

1. Rendering Class — Stylized 3D, Flat 2D, Minimalist Line Art, Textured 2D, Mixed Media, Photoreal
2. World Type — Stylized Real-World, Abstract Concept Space, Data/Presentation Space, Fictional Metaphor Universe
3. Character Strategy — None, Mascot-Led, Single Narrator, Single Protagonist Arc, Ensemble Cast
4. Narrative Structure — Direct Explanation, Step-by-Step, Problem-Solution, Analogy, Myth-Busting, etc.
5. Visual Abstraction Index — 1 (Photorealistic) to 5 (Maximum Abstraction)
6. Visual Density — Minimal, Sparse, Moderate, High
7. Camera/Editing Language — Cinematic, Social Vertical Punch, Presentation Deck, Static Slides, etc.
8. Tonal Positioning — Institutional, Corporate Professional, Gen Z Social, Child-Friendly, Dark Editorial

License

MIT