multi-agent-system

Multi-Agent AI Orchestration System

This project is a multi-agent AI system designed to handle user requests by coordinating between different specialized AI agents. It demonstrates a hierarchical architecture, intent recognition, and robust error handling.

🏛️ Architecture

This system uses a three-level hierarchical agent architecture:

• Level 1 (Super Agent): The central orchestrator. [cite_start]It receives user requests, delegates tasks to mid-level agents, and returns the final response. [cite: 12]
• Level 2 (Mid-Level Agents):
  • [cite_start]Intent Analysis Agent: Interprets the user's natural language query to determine their goal (intent) and key information (entity). [cite: 14]
  • [cite_start]Error Monitoring Agent: Logs any errors that occur within the system to a file for later review. [cite: 15]
• [cite_start]Level 3 (Task Agents): Worker agents that perform specific, predefined operations. [cite: 16]

Communication Flow

The communication flows from top to bottom and back up.

graph TD
    User -->|Request| A[L1: Super Agent];
    A -->|"What's the intent?"| B[L2: Intent Agent];
    B -->|"Intent: get_weather<br/>Entity: London"| A;
    A -->|"Get weather for London"| C[L3: Weather Agent];
    C -->|API Call| D[External Weather API];
    C -.->|Logs Error| E[L2: Error Monitoring Agent];
    D -->|Weather Data| C;
    C -->|Weather Info| A;
    A -->|Final Response| User;

(This is a Mermaid diagram. It will render automatically on platforms like GitHub.)

✨ Features Implemented

• [cite_start]Hierarchical Agent System: A three-level architecture for clear separation of concerns. [cite: 23]
• [cite_start]Natural Language Intent Recognition: Uses the Groq LLM to understand user queries and map them to appropriate agents. [cite: 27]
• Dynamic Task Routing: The Super Agent dynamically calls the correct task agent based on the recognized intent.
• [cite_start]Centralized Error Logging: A dedicated Error Monitoring Agent captures and logs failures from other agents. [cite: 31]
• Expandable Agent Registration: New task agents can be added easily using a simple decorator, requiring no changes to the core system logic.
• Implemented Task Agents:
  • [cite_start]get_weather: Retrieves current weather for a specified location. [cite: 17]
  • [cite_start]get_news: Fetches the latest news headlines for a given topic. [cite: 18]

🚀 Setup and Installation

1. Clone the repository:

   git clone <your-repo-url>
   cd multi-agent-system
   

2. Create and activate a virtual environment:

   python -m venv venv
   # On Windows: venv\Scripts\activate
   # On macOS/Linux: source venv/bin/activate
   

3. Install the required dependencies:

   pip install -r requirements.txt
   

4. Configure API keys:

   • Create a file named .env in the root directory.
   • Add your API keys to the .env file:

     GROQ_API_KEY="YOUR_GROQ_API_KEY"
     WEATHER_API_KEY="YOUR_OPENWEATHERMAP_API_KEY"
     NEWS_API_KEY="YOUR_NEWS_API_KEY"
     

5. Run the application:

   python main.py
   

🎬 Demo

(You will add a link to your demo video here later)