Multi-Agent AI System - Task & Schedule Manager 🤖

https://sivas-multi-agent-api.onrender.com/docs

Submission for Hack2Skill Hackathon (Track 3)

This project is a multi-agent AI system designed to help users manage tasks, schedules, and information by interacting with multiple tools and data sources. It is deployed as a fully functional API using FastAPI.

🌟 Live Demo & Links

• Live API URL: [Insert your Render URL here, e.g., https://my-agent.onrender.com/docs]
• Demo Video: [Insert link to your YouTube/Drive demo video here]

🎯 Core Requirements Met

1. Primary Agent Coordination: A Primary Orchestrator agent receives user prompts and routes them to the appropriate sub-agents.
2. Sub-Agents: Features specialized sub-agents (Database Manager and Calendar Manager) to execute distinct workflows.
3. Structured Database: Integrates an SQLite database (hackathon.db) to store, retrieve, and track generated tasks.
4. Tool Integration (MCP-style): Agents are equipped with specific functional tools to mock external interactions (e.g., calendar scheduling).
5. API-Based Deployment: The entire system is wrapped in a FastAPI application, allowing external platforms to trigger multi-step workflows via standard HTTP POST requests.

🛠️ Tech Stack

• Language: Python 3
• API Framework: FastAPI & Uvicorn
• Database: SQLite
• Agent Logic: Custom LLM routing logic / Pydantic

🚀 How to Run Locally

1. Clone the repository:

   git clone https://github.com/your-username/your-repo-name.git
   cd your-repo-name

2. Install dependencies: Make sure you have Python installed, then install the required packages using pip:

   pip install -r requirements.txt

   (Requirements included: fastapi, uvicorn, pydantic, requests)

3. Set up Environment Variables (Optional): The application uses OpenAI's API to power the primary agent router. If an API key is not provided, it falls back to a mock agent behavior for testing.

   # On Windows (PowerShell)
   $env:OPENAI_API_KEY="your-api-key-here"

   # On Linux/macOS
   export OPENAI_API_KEY="your-api-key-here"

4. Start the API server: Run the application using Uvicorn from the project directory:

   uvicorn main:app --reload

5. Test the API: Open your browser and navigate to http://127.0.0.1:8000/docs to use the built-in Swagger UI.

📡 API Endpoints

POST /execute_task

This is the primary endpoint to interact with the multi-agent system.

Request Body (JSON):

{
  "user_request": "Schedule a hackathon review meeting for tomorrow, and save a task to submit the project files."
}

Expected Response (JSON):

{
  "status": "success",
  "original_message": "Schedule a hackathon review meeting for tomorrow, and save a task to submit the project files.",
  "agent_response": "Success: Saved 'Demo Video' to database. | Success: Scheduled event 'Hackathon Review'."
}

🏗️ System Architecture

1. User Request -> Sent via POST to /execute_task.
2. Primary Agent (Router) -> Analyzes the intent of the prompt.
3. Delegation -> Routes to:
   • Database Sub-Agent: Extracts task data and writes it to SQLite.
   • Calendar Sub-Agent: Extracts event data and processes the schedule.
4. Response Aggregation -> The primary agent combines the tools' outputs and returns a final success confirmation to the user.