Argus

Argus

A MCP (Model Context Protocol) server for analyzing videos and extracting UI/animation design specifications using OpenAI's Vision API.

Install

# 1. Install ffmpeg (required)
brew install ffmpeg

# 2. Install Argus
curl -fsSL https://raw.githubusercontent.com/jamesrochabrun/Argus/main/install.sh | sh

The script downloads the binary and shows configuration instructions.

Configure Claude Code

Add to ~/.claude.json:

{
  "mcpServers": {
    "argus": {
      "type": "stdio",
      "command": "~/.local/bin/argus",
      "args": ["mcp"],
      "env": {
        "OPENAI_API_KEY": "your-openai-api-key"
      }
    }
  }
}

Get your API key at: https://platform.openai.com/api-keys

Restart Claude Code to use Argus.

---

Tools

Tool	Purpose	Cost
analyze_video	Describe video content (UI, animations, design elements)	~$0.001-0.003
design_from_video	Extract animation specs for implementation in any framework	~$0.003-0.01

analyze_video

Describes what's happening in a video.

Analyze this video: /path/to/recording.mov

design_from_video

Extracts animation timing, curves, and choreography into a framework-agnostic spec.

Parameters:

• video_path: Path to video file
• mode: quick or high_detail
• focus_hint (optional): Element to focus on

Extract animation specs from /path/to/animation.mov using quick mode

Output: JSON spec with elements, keyframes, curves - ready for SwiftUI, React, Flutter, etc.

---

Build from Source

For contributors:

git clone https://github.com/jamesrochabrun/Argus.git
cd Argus
swift build -c release
.build/release/argus --setup

---

License

MIT