🤖 Copilot LLM Chat Studio

Multi-LLM Collaborative Chat Interface - Select 2-3 local Ollama models to work together on dev projects

🎯 Vision

Imagine having multiple AI coding specialists working together in the same chat. Select your best local Ollama models, give them a project, and watch them collaborate, plan, code, test, and build together.

You: "Hey team, I want to build a REST API with authentication"

🧠 DeepSeek-Coder: "I'll design the architecture and core endpoints..."
🔧 Qwen2.5-Coder: "I'll implement the auth middleware and JWT handling..."
🧪 CodeGemma: "I'll write the test suite and security checks..."

[Models read and build upon each other's responses]

✨ Features

• 🎚️ Model Selector - Dropdown to pick 2-3 Ollama models from your 67+ local arsenal
• 💬 Unified Chat - Single interface where all selected models participate
• 🔄 Inter-Model Communication - Models see and respond to each other's outputs
• 📁 Project Context - Load codebase files for context-aware responses
• ▶️ Code Execution - Run generated code snippets directly
• 💾 Export/Save - Save conversations and generated code
• 🌙 Dark Mode - Easy on the eyes for those late night coding sessions

🏗️ Architecture

┌─────────────────────────────────────────────────────────────┐
│                    Electron Desktop App                      │
├─────────────────────────────────────────────────────────────┤
│  ┌─────────────────┐  ┌─────────────────┐  ┌─────────────┐ │
│  │  Model Selector │  │   Chat Panel    │  │  Code View  │ │
│  │   (2-3 models)  │  │  (all models)   │  │  (Monaco)   │ │
│  └────────┬────────┘  └────────┬────────┘  └──────┬──────┘ │
│           │                    │                   │        │
│  ┌────────┴────────────────────┴───────────────────┴──────┐ │
│  │              Conversation Orchestrator                  │ │
│  │   • Turn management  • Context building  • Streaming   │ │
│  └────────────────────────┬────────────────────────────────┘ │
└───────────────────────────┼─────────────────────────────────┘
                            │
                   ┌────────▼────────┐
                   │   Ollama API    │
                   │ localhost:11434 │
                   │   67+ Models    │
                   └─────────────────┘

🚀 Quick Start

Prerequisites

• Node.js 18+
• Ollama running locally
• At least 2-3 models pulled (e.g., deepseek-coder, qwen2.5-coder, codellama)

Installation

# Clone the repository
git clone https://github.com/VisionaryArchitects/Copilot_LLM_Chat_Studio.git
cd Copilot_LLM_Chat_Studio

# Install dependencies
npm install

# Start development
npm run dev

# Or build and run
npm run build
npm start

Recommended Models

Model	Specialty	Size
deepseek-coder:33b	Advanced code generation	33B
qwen2.5-coder:32b	Multi-language coding	32B
codellama:34b	Code completion & infilling	34B
codegemma:7b	Fast code assistant	7B
starcoder2:15b	Code generation	15B

📖 Usage

1. Select Your Team

Use the model selector to pick 2-3 models. Mix specialists for best results!

2. Start a Conversation

Type your project idea or coding task. All selected models will see your message.

3. Watch Them Collaborate

Models read each other's responses and build upon them. You can:

• Let them auto-collaborate
• Direct specific models with @model-name
• Request specific actions: "plan", "code", "review", "test"

4. Execute & Export

• Run code snippets directly
• Copy generated code
• Export full conversations

🔧 Configuration

Create a .env file:

OLLAMA_API_URL=http://localhost:11434
MAX_CONTEXT_LENGTH=8192
DEFAULT_MODELS=deepseek-coder:33b,qwen2.5-coder:32b

🛠️ Development

# Run in development mode
npm run dev

# Lint code
npm run lint

# Run tests
npm test

# Package for distribution
npm run package

📁 Project Structure

Copilot_LLM_Chat_Studio/
├── src/
│   ├── main/           # Electron main process
│   ├── renderer/       # React frontend
│   │   ├── components/ # UI components
│   │   ├── hooks/      # Custom React hooks
│   │   ├── contexts/   # React contexts
│   │   ├── utils/      # Utility functions
│   │   └── styles/     # CSS/Tailwind
│   └── shared/         # Shared types & constants
├── server/             # Local Express server (optional)
├── docs/               # Documentation & assets
└── package.json

🤝 Contributing

We welcome contributions! Please see CONTRIBUTING.md for guidelines.

📜 License

MIT License - see LICENSE for details.

🏢 Organization

Visionary Architects - AI Development Ecosystem

---

Built with ❤️ for the AI-first development community