AutoGen Multi-Agent System

Production-quality multi-agent collaboration system using Microsoft's AutoGen framework with professional dark theme dashboard.

🚀 Live Demo

View in AI Portfolio Dashboard | Direct Access

Part of Nihal's AI Portfolio - Unified dashboard featuring 5 cutting-edge AI services

Features

4 Specialized Agents:

• Researcher: Web search and information gathering
• Coder: Python code generation and execution
• Reviewer: Quality control and feedback
• Synthesizer: Final analysis and recommendations

Production Dashboard:

• Professional dark theme interface
• Real-time agent collaboration streaming
• Process transparency with tool execution visibility
• Timestamped messages with color-coded agents

Comprehensive Tools:

• Web search via Google Serper API
• Python code execution (numpy, pandas, matplotlib, scikit-learn)
• File management operations
• Safe sandboxed environment

Quick Start

# Install dependencies
python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt

# Setup environment
cp .env.example .env
# Add your OPENAI_API_KEY and SERPER_API_KEY to .env

# Run dashboard
python dashboard.py
# Open https://unharmable-threadlike-ruth.ngrok-free.dev:8000

Available Demos

1. dashboard.py - Professional web UI with 4-agent system (PRIMARY)
2. multi_agent.py - CLI version with 4 agents and tools
3. demo.py - Simple 2-agent collaboration
4. autogen_core_demo.py - RoutedAgent and Runtime patterns
5. app.py - Code review team example

Technical Stack

• AutoGen 0.4.0 (multi-agent framework)
• FastAPI (web backend)
• Server-Sent Events (real-time streaming)
• OpenAI GPT-4 (agent intelligence)
• Google Serper (web search)
• Python scientific libraries (numpy, pandas, matplotlib, scikit-learn)

Architecture

Dashboard (FastAPI + SSE)
├── 4 Specialized Agents
│   ├── Researcher (web search tool)
│   ├── Coder (code execution tool)
│   ├── Reviewer (quality control)
│   └── Synthesizer (final summary)
├── RoundRobinGroupChat (team coordination)
├── Tool Integration (search, execute, files)
└── Real-time Streaming (process transparency)

Example Tasks

• "Explain quantum computing and implement a qubit simulation"
• "Research machine learning and implement linear regression"
• "Analyze blockchain technology and create basic implementation"

Requirements

• Python 3.10+
• OpenAI API key
• Google Serper API key (optional, for web search)

Production Quality

• Clean modular code structure
• Error handling and graceful degradation
• Professional UI/UX design
• Comprehensive documentation
• Tested end-to-end

License

MIT License

---

Production-ready multi-agent AI system with professional dashboard and comprehensive tool integration.