Ollama SQLite Memory Chatbot

Local AI chatbot using Ollama with persistent conversation memory stored in SQLite.

Designed for low-resource devices (CPU-only laptops) without GPU.

---

🚀 Features

• Local LLM using Ollama
• Persistent chat memory with SQLite
• Session-based conversations
• Lightweight design (CPU friendly)
• Simple CLI interface

---

🧠 Architecture

User → Python CLI → SQLite Memory → Ollama → Response

---

⚙️ Requirements

Install dependencies:

pip install ollama

Make sure Ollama is installed and running:

https://ollama.com

Pull a lightweight model:

ollama pull llama3.2:1b

📦 Configuration

Edit config.json:

{
  "model": "llama3.2:1b",
  "system_prompt": "You are Jajang, a helpful AI assistant. Answer clearly and concisely.",
  "stream": false
}

▶️ How to Run

Run chatbot:

python app.py

💬 Example Usage

You: halo selamat malam
AI: Selamat malam! Ada yang bisa saya bantu?

You: nama kamu siapa
AI: Saya Jajang, asisten AI kamu.

🗄️ Memory System

Chat history is stored in SQLite database:

• Persistent across sessions
• Session-based isolation
• Automatically loaded into model context

⚠️ Limitations

• Uses small model (llama3.2:1b)
• Running on CPU only may cause slower responses
• Possible hallucination due to model size
• Limited context window

💡 Notes

This project is designed for:

• Learning LLM integration
• Local AI experimentation
• Lightweight chatbot systems
• Embedded / low-resource AI applications

Not intended for production-scale deployment.

📄 License

MIT License