A lightweight CLI-based AI assistant framework that supports memory, session saving, persona-driven response generation, and OpenAI integration.

🎯 Features

• CLI-based interface for interacting with your AI persona
• Persona system with customizable tone, goals, and constraints
• Session logging (stored with timestamped filenames)
• Offline mode for CLI + command testing only (stub returns same default message across personas)
• Realistic AI conversations available via OpenAI integration
• Modular design (easy to extend or plug into other tools)

📦 Project Structure

• main.py – Main entrypoint (interactive chat loop)
• memory.py – Tracks multi-turn dialogue context
• session_saver.py – Logs conversations with timestamps
• prompt_builder.py – Constructs persona-aware prompts
• response_handler.py – Stub/OpenAI response engine
• persona_config.json – Persona definitions
• templates/ – Prompt templates by persona

🧠 Memory + Session Logging

All dialogue is tracked and saved when the user exits via exit or quit. Session logs are timestamped and saved locally.

⌨️ Supported CLI Commands

/save           Save current session
/delete         Delete current persona
/traits         View traits of current persona
/history        View full conversation history
/reset          Reset memory for current persona
/help           Show command list
exit | quit     End session and save

🔧 To Run

python main.py

Offline mode will launch with a static preset message. Use this for testing CLI commands and flows.

🌐 OpenAI Integration (Recommended for Full Features)

To enable full AI responses and personality simulation:

1. Create a file named config.py in the root directory
2. Add your OpenAI API key to it:

OPENAI_API_KEY = "your-openai-api-key-here"

3. In response_handler.py, set:

USE_OPENAI = True

⚠️ Stub mode returns a single static message. OpenAI integration is required for realistic persona behavior.

👤 About the Author

Justin Lane
🔗 GitHub: @aiwithjusl
🔗 LinkedIn: Justin Lane
📬 Email: aiwithjusl.dev@gmail.com

📄 License

This project is licensed under the MIT License.