Multi-LLM Application

A powerful and flexible application for managing and interacting with multiple Large Language Models (LLMs) through a unified interface.

Features

• 🤖 Multi-Model Support: Seamlessly integrate with various LLM providers:

  • Local models via LM Studio
  • OpenAI models (GPT-3.5, GPT-4)
  • Anthropic's Claude (Sonnet, Haiku)
  • Deepseek models (Chat, Coder)
  • Custom model implementations

• 💬 Advanced Chat Interface:

  • Personalized chat agent with customizable name and personality
  • Multi-session management
  • Context-aware conversations
  • File attachments support
  • Real-time responses

• 📊 Vector Database Integration:

  • Document embeddings with Weaviate
  • Semantic search capabilities
  • Efficient document management
  • Automatic indexing

• 🤖 Agent System:

  • Create and manage AI agents
  • Tool integration
  • Task automation
  • Status monitoring

• 🛠️ Tool Management:

  • Custom tool creation
  • API integrations
  • Function calling
  • CLI command execution

• 🧠 Model Context Protocol (MCP):

  • Context management
  • Cross-model communication
  • Context merging
  • Response tracking

• 💻 CLI Interface:

  • Command-line control
  • System monitoring
  • Agent management
  • Quick actions

Getting Started

Prerequisites

• Node.js 18 or higher
• npm or yarn
• A modern web browser
• ChromaDB v0.6.0 or higher

ChromaDB v0.6.0 Migration

The application has been updated to support ChromaDB v0.6.0, which includes several important changes:

1. Collection Listing: ChromaDB now returns only collection names from list_collections() instead of full collection objects.
2. Storage Backend: Uses SQLite by default instead of DuckDB+Parquet.
3. Client Initialization: Uses the new PersistentClient class:

   client = chromadb.PersistentClient(
       path="./chroma_data",
       settings=Settings(
           anonymized_telemetry=False,
           allow_reset=True,
           is_persistent=True
       )
   )

These changes improve reliability and performance while maintaining compatibility with existing data.

Installation

1. Clone the repository:

   git clone https://github.com/yourusername/multi-llm-app.git
   cd multi-llm-app

2. Run the installation script:

   install.bat

   The installation script provides several options:

   • Full Installation (All Components)
   • Frontend Only (npm packages)
   • Python Environment Only
   • ChromaDB Setup Only
   • Weaviate Setup Only

   Choose the appropriate option based on your needs. For first-time setup, select "Full Installation".

3. Configure environment variables:

   • Copy .env.example to .env
   • Update the configuration values in .env

4. Start the application:

   start.bat

   This will start both the ChromaDB server and the frontend development server.

Note: If you encounter any issues during installation, you can run individual components of the installation process by selecting the appropriate option in install.bat.

Configuration

1. Navigate to the Settings page
2. Configure your LLM providers:
   • LM Studio URL (for local models)
   • OpenAI API key and model selection
   • Claude API key and model selection (Sonnet/Haiku)
   • Deepseek API key and model selection (Chat/Coder)
3. Set your default provider for the chat agent
4. Optional: Set up Weaviate for document embeddings

Chat Agent Configuration

The chat agent can be customized through the Chat Configuration page:

1. Name: Set a custom name that the agent will respond to

2. Personality:

   • Traits: Define characteristics (e.g., friendly, professional)
   • Tone: Choose from professional, casual, friendly, or formal
   • Style: Select concise, detailed, technical, or simple
   • Constraints: Add specific behavioral rules

3. Language Model:

   • Provider: Use the default from settings or choose another
   • Model: Select from available models for the chosen provider
   • Parameters: Adjust temperature and max tokens

The chat agent's configuration syncs with your settings:

• The default provider from settings is automatically selected
• Changing the provider in chat config updates the default in settings
• Only enabled providers from settings are available

Architecture

The application is built with modern web technologies:

• Frontend: React, TypeScript, Tailwind CSS
• State Management: Zustand
• Routing: React Router
• Vector Database: Weaviate
• Embedding Storage: ChromaDB
• UI Components: Custom components with Lucide icons

Components

• Layout: Main application structure
• ChatInput/Message: Chat interface components
• AgentCard/Form: Agent management
• ToolCard/Form: Tool configuration
• DocumentUploader/List: Document handling
• Modal: Reusable modal component
• StatusCard/StatsCard: Dashboard components

Pages

• /: Chat interface
• /embeddings: Document embeddings
• /dashboard: System overview
• /agents: Agent management
• /tools: Tool configuration
• /mcp: Model Context Protocol
• /cli: Command-line interface
• /settings: System configuration

Development

Project Structure

src/
  ├── components/     # Reusable UI components
  ├── pages/         # Application pages
  ├── lib/           # Utility functions and services
  ├── store/         # State management
  ├── types/         # TypeScript definitions
  └── main.tsx       # Application entry point

Building

npm run build

Testing

npm run test

Contributing

1. Fork the repository
2. Create a feature branch
3. Commit your changes
4. Push to the branch
5. Create a Pull Request

Git Configuration

The repository includes comprehensive .gitignore settings for:

• Database files and data (chroma_data/, chroma_backup/)
• Python-specific files (pycache/, *.pyc)
• Development and build files (dist/, build/)
• Environment and configuration files (.env)
• Local development files (.vite/, .cache/)
• System and temporary files

License

This project is licensed under the MIT License - see the LICENSE file for details.

Acknowledgments

• Icons by Lucide
• UI inspiration from various modern web applications
• Community feedback and contributions

Support

For support, please:

1. Check the documentation
2. Search existing issues
3. Create a new issue if needed

---

Built with ❤️ by [Your Name/Organization]