SSMIS - Local LLM Chat Interface

A powerful local LLM chat application with support for Ollama and LM Studio models, featuring a modern PyQt5 interface, TTS integration, and GPU acceleration.

Features

• 🤖 Multi-Model Support: Seamlessly switch between Ollama and LM Studio models without restart
• 💬 Chat Interface: Modern dark-themed chat window with message history
• 🔊 TTS Integration: OpenAI-compatible Text-to-Speech for voice output
• 💾 Persistent Context: MCP memory server connector for maintaining conversation context
• 🖥️ System Tray: Minimize to system tray for background operation
• ⚡ GPU Acceleration: CUDA 12.1/12.8 support for faster inference
• 🔧 Configurable: Comprehensive settings for ports, models, and preferences
• 📁 Chat History: Save and browse previous conversations

Requirements

• Python 3.11+
• PyQt5
• aiohttp
• Optional: NVIDIA GPU with CUDA 12.1 or 12.8 for GPU acceleration

Installation

Quick Setup

# Clone the repository
git clone https://github.com/tevasauksinis1-ctrl/SSMIS.git
cd SSMIS

# Run the setup script (creates isolated Python 3.11 environment)
chmod +x setup_env.sh
./setup_env.sh

Manual Installation

# Create virtual environment
python3.11 -m venv venv
source venv/bin/activate

# Install dependencies
pip install -r requirements.txt

# For CUDA 12.1 support
pip install torch --index-url https://download.pytorch.org/whl/cu121

# For CUDA 12.8 support
pip install torch --index-url https://download.pytorch.org/whl/cu128

Usage

Starting the Application

# Using the launcher
~/.llm_chat/run.sh

# Or manually
python -m llm_chat

# Start minimized to system tray
python -m llm_chat --minimized

# Enable debug logging
python -m llm_chat --debug

# Use custom config file
python -m llm_chat --config /path/to/config.json

Prerequisites

Before using LLM Chat, ensure you have one of the following running:

1. Ollama (default port 11434):

   # Install Ollama
   curl -fsSL https://ollama.com/install.sh | sh
   
   # Pull a model
   ollama pull llama2
   
   # Start Ollama
   ollama serve

2. LM Studio (default port 1234):

   • Download from lmstudio.ai
   • Load a model and start the local server

Optional: TTS Service

For text-to-speech functionality, you need an OpenAI-compatible TTS server:

# Example using xtts-api-server
pip install xtts-api-server
xtts-api-server --port 8000

Configuration

Configuration is stored at ~/.llm_chat/config.json. You can also access settings through the application's Settings dialog (Edit → Settings).

Configuration Options

Setting	Default	Description
ports.ollama	11434	Ollama API port
ports.lm_studio	1234	LM Studio API port
ports.tts	8000	TTS service port
ports.mcp_memory	3100	MCP memory server port
model.default_provider	ollama	Default model provider
model.temperature	0.7	Sampling temperature
model.max_tokens	2048	Maximum response tokens
tts.enabled	false	Enable TTS
tts.voice	alloy	TTS voice
mcp.enabled	true	Enable MCP memory
cuda.enabled	true	Enable CUDA
cuda.cuda_version	12.1	CUDA version

Architecture

llm_chat/
├── __init__.py          # Package initialization
├── __main__.py          # Application entry point
├── core/
│   ├── chat_client.py   # Main chat client
│   ├── model_manager.py # Model switching
│   └── history_manager.py # Chat history
├── gui/
│   ├── main_window.py   # Main application window
│   ├── chat_widget.py   # Chat display widget
│   ├── system_tray.py   # System tray manager
│   └── dialogs/         # Settings, model, about dialogs
├── services/
│   ├── ollama_client.py # Ollama API client
│   ├── lm_studio_client.py # LM Studio client
│   ├── tts_service.py   # TTS integration
│   └── mcp_connector.py # MCP memory connector
├── config/
│   └── settings.py      # Application settings
└── utils/
    ├── config_manager.py # Configuration management
    └── logger.py        # Logging utilities

Keyboard Shortcuts

Shortcut	Action
Ctrl+N	New Chat
Ctrl+L	Clear Chat
Ctrl+E	Export Chat
Ctrl+M	Select Model
Ctrl+H	Toggle History Panel
F5	Refresh Models
Ctrl+,	Settings

Development

# Install development dependencies
pip install pytest pytest-asyncio black mypy

# Run tests
pytest

# Format code
black llm_chat/

# Type checking
mypy llm_chat/

License

Apache License 2.0 - See LICENSE for details.

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.