CHAT-AI

A modular, privacy-aware, and AI-powered chat system with both terminal and web interfaces.
Supports local and server-based AI, context-aware replies, privacy controls, and robust logging.

---

Features

• AI Reply Prediction: Local rule-based or server LLM (Ollama, Flask, etc.)
• Privacy Controls: Configurable privacy guard, encrypted logs
• Context Awareness: Recent message context for smarter AI
• Web API & Demo: FastAPI backend and browser chat demo
• Logging: Per-user logs, optional encryption
• Extensible: Modular codebase for easy upgrades

---

Project Structure

ai/
    local_small_llm.py         # Local AI/tone logic
    reply_predictor.py         # AI reply prediction logic
client/
    advanced_tone_model.py     # Tone analysis/generation
    context_vectorizer.py      # Context embedding
    main_client.py             # Terminal chat client
    privacy_guard.py           # Privacy settings/logic
server/
    context_manager.py         # Tracks conversation context
    llm_interface.py           # Unified LLM API (local/remote)
    main_server.py             # Multi-client chat server
    logs/
        alice_logs.json        # Example user log
        message_logger.py      # Logging utility
shared/
    constants.py               # Shared constants
    encryption_utils.py        # Encryption helpers
    utils.py                   # Shared utilities
tests/
    test_client_server.py      # Integration tests
    test_privacy_guard.py      # Privacy guard tests
    test_reply_predictor.py    # AI reply tests
web/
    web_api.py                 # FastAPI REST API
    static
static.index.html
static.custom.css
static.app.js              # Browser chat demo

---

Getting Started

1. Install dependencies

pip install -r requirements.txt

2. Start the server

python server/main_server.py

3. (Optional) Start the FastAPI web API

uvicorn web.web_api:app --reload

4. Run the terminal client

python client/main_client.py

5. (Optional) Open the web demo

Open web/web_demo.html in your browser.
If running the API on a different host/port, adjust the fetch URL in the HTML.

---

Configuration

Edit client/config.json (if present) or adjust settings in privacy_guard.py for:

• AI backend selection (local/server)
• Privacy features
• Log encryption

---

Testing

Run all tests with:

python -m unittest discover tests

---

Extending

• Add new AI models in ai/
• Add more endpoints or web features in web/
• Expand privacy and logging as needed

---

License

MIT License (or your choice)