chatlite 0.1.7

ai powered chatapp for browsing and search

ChatLite 🤖

A lightweight, extensible chat application framework for building AI-powered chat interfaces. ChatLite provides an easy-to-use platform for integrating various language models with web-based chat applications.

✨ Features

• 🔄 Real-time WebSocket communication
• 🎯 Multi-model support (Llama, Qwen, etc.)
• 🌐 Web search integration
• 🎨 Customizable UI with modern design
• 🔌 Plugin architecture for easy extensions
• 💬 Chat history management
• 🎭 Multiple agent types support
• 📱 Responsive design

🚀 Quick Start

Installation

pip install chatlite

Basic Usage

import chatlite

# Start a simple chat server with Llama 3.2
chatlite.local_llama3p2()

# Or use Qwen 2.5
chatlite.local_qwen2p5()

# Custom configuration
server = chatlite.create_server(
    model_type="local",
    model_name="llama3.2:latest",
    temperature=0.7,
    max_tokens=4000
)
server.run()

Pre-configured Models

ChatLite comes with several pre-configured models:

# Use different models directly
from chatlite import mistral_7b_v3, mixtral_8x7b, qwen_72b

# Start Mistral 7B server
mistral_7b_v3()

# Start Mixtral 8x7B server
mixtral_8x7b()

# Start Qwen 72B server
qwen_72b()

💻 Frontend Integration

ChatLite includes a Flutter-based frontend that can be easily customized. Here's a basic example of connecting to the ChatLite server:

final channel = WebSocketChannel.connect(
  Uri.parse('ws://localhost:8143/ws/$clientId'),
);

// Send message
channel.sink.add(json.encode({
  'message': 'Hello!',
  'model': 'llama3.2:latest',
  'system_prompt': 'You are a helpful assistant',
  'agent_type': 'WebSearchAgent',
  'is_websearch_chat': true
}));

// Listen for responses
channel.stream.listen(
  (message) {
    final data = jsonDecode(message);
    if (data['type'] == 'stream') {
      print(data['message']);
    }
  },
  onError: (error) => print('Error: $error'),
  onDone: () => print('Connection closed'),
);

🔧 Configuration

ChatLite supports various configuration options:

from chatlite import create_server

server = create_server(
    model_type="local",          # local, huggingface, etc.
    model_name="llama3.2:latest",
    api_key="your-api-key",      # if needed
    temperature=0.7,             # model temperature
    max_tokens=4000,             # max response length
    base_url="http://localhost:11434/v1",  # model API endpoint
)

🧩 Available Agents

ChatLite supports different agent types for specialized tasks:

• WebSearchAgent: Internet-enabled chat with web search capabilities
• RawWebSearchAgent: Direct web search results without summarization
• EmailAssistantFeature: Email composition and analysis
• DefaultChatFeature: Standard chat functionality

Example usage:

# Client-side configuration
message_data = {
    "message": "What's the latest news about AI?",
    "model": "llama3.2:latest",
    "agent_type": "WebSearchAgent",
    "is_websearch_chat": True
}

🎨 UI Customization

The included Flutter frontend supports extensive customization:

ThemeData(
  brightness: Brightness.dark,
  scaffoldBackgroundColor: const Color(0xFF1C1C1E),
  primaryColor: const Color(0xFF1C1C1E),
  colorScheme: const ColorScheme.dark(
    primary: Color(0xFFFF7762),
    secondary: Color(0xFFFF7762),
  ),
)

📦 Project Structure

chatlite/
├── __init__.py          # Main package initialization
├── core/               # Core functionality
│   ├── config.py       # Configuration handling
│   ├── model_service.py # Model interaction
│   └── features/       # Feature implementations
├── ui/                 # Flutter frontend
└── examples/           # Usage examples

🤝 Contributing

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

1. Fork the repository
2. Create your feature branch (git checkout -b feature/AmazingFeature)
3. Commit your changes (git commit -m 'Add some AmazingFeature')
4. Push to the branch (git push origin feature/AmazingFeature)
5. Open a Pull Request

🙏 Acknowledgments

• Built with FastAPI and Flutter
• Inspired by modern chat applications
• Uses various open-source language models