🧠 AI Planning Assistant

An AI Agent that helps you organize your study schedule, LeetCode practice, and daily habits — built with Python, MCP, and Starlette.

---

🚀 Features

• Fetch school schedule from API.
• Plan smart LeetCode practice sessions.
• Maintain habits using RAG (Retrieval-Augmented Generation).
• Real-time streaming responses with SSE.

---

🛠 Installation

1. Clone the repository & create a virtual environment

git clone https://github.com/nax-handle/mcp.git
cd mcp

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

2. Install dependencies

pip install -r requirements.txt

---

⚙️ Environment Variables

Create a .env file in the root folder:

OPENAI_API_KEY=your-openai-api-key
DATABASE_URL=your-database-url
DEBUG=true

---

🧪 Running the App

Development Mode (auto-reload):

mcp run server.py

By default, the app will run on: http://localhost:8000

---

📂 Project Structure

.
├── app/
│   ├── agents/              # MCP Agents
│   ├── services/            # External API calls (School, LeetCode, Habit...)
│   ├── schemas/             # Pydantic models
│   ├── main.py              # FastAPI/Starlette app
├── .env
├── requirements.txt
└── README.md

---

⚡ Example API Request

curl http://localhost:8000/schedule?user_id=123

The response will be in application/json or streamed via SSE if requested.

---

🤖 MCP Agent Workflow

flowchart TD
    A[User Request] --> B{Agent}
    B --> C[Get School Schedule]
    B --> D[Get LeetCode Tasks]
    B --> E[RAG: Get Habit Data]
    C & D & E --> F[Planner]
    F --> G[Return Optimized Plan]

Loading

---

📌 Notes

• The project uses MCP (Multi-agent Control Plane) to manage complex logic in a modular fashion.
• It's easy to extend and integrate with other APIs (Notion, Google Calendar, Zalo Reminder, etc.).