Nexus Ai

Nexus

A native desktop AI chat application with multi-provider support and MCP integration. Built with Tauri v2, React 19, and TypeScript.

📸 Screenshots

✨ Features

Multi-Provider LLM Support

• Google Gemini - 1.5 Flash/Pro, 2.0 Flash, 2.5 Flash/Pro, 3.0 Pro & Deep Think
• OpenAI - GPT-4o, GPT-4o Mini, GPT-4 Turbo, DALL-E 3
• Groq - Llama 3.3 70B, Llama 3.1 8B, Mixtral 8x7B, Llama 3.2 Vision

Multimodal Capabilities

• 📸 Vision - Analyze images with vision-capable models
• 🎤 Audio - Voice input with audio transcription
• 🎨 Image Generation - Create images with DALL-E 3 and Gemini

MCP Tool Integration

• Full Model Context Protocol support
• Connect to MCP servers via stdio or HTTP/SSE transports
• Execute tools directly from chat conversations

Native Desktop Experience

• ⌨️ Command Palette - Quick access to all actions (Cmd+K)
• 💬 Conversation History - Persistent chat history with search
• 📤 Export - Save conversations as Markdown or JSON
• 🎨 Themes - Light, dark, and system themes
• 🔄 Auto Updates - Built-in updater with GitHub releases

📦 Installation

Prerequisites

• Node.js (v18+)
• Rust (latest stable)
• Xcode Command Line Tools (macOS)

Development Setup

# Clone the repository
git clone https://github.com/navjotdhanawat/nexus.git
cd nexus

# Install dependencies
npm install

# Start development server
npm run tauri:dev

Building for Production

# Build the application
npm run tauri:build

# The .app and .dmg will be in src-tauri/target/release/bundle/

🚀 Quick Start

1. Launch the app and open Preferences (Cmd+,)
2. Add your API keys in the API Keys section:

• Google AI API key for Gemini models
• OpenAI API key for GPT and DALL-E
• Groq API key for Llama and Mixtral

1. Start chatting! Select a model and send your first message

Keyboard Shortcuts

| Shortcut | Action | |----------|--------| | Cmd+K | Open Command Palette | | Cmd+, | Open Preferences | | Cmd+1 | Toggle Left Sidebar | | Cmd+2 | Toggle Right Sidebar | | Cmd+N | New Conversation | | Enter | Send Message | | Shift+Enter | New Line |

🔧 Configuration

MCP Servers

Configure MCP servers in Preferences → MCP Servers:

{
  "name": "My MCP Server",
  "transport": "stdio",
  "command": "npx",
  "args": ["-y", "@modelcontextprotocol/server-filesystem", "/path/to/directory"]
}

Environment Variables

For development, you can set API keys via environment variables:

VITE_GOOGLE_API_KEY=your-google-api-key
VITE_OPENAI_API_KEY=your-openai-api-key
VITE_GROQ_API_KEY=your-groq-api-key

🏗️ Architecture

├── src/                    # React frontend
│   ├── components/         # UI components
│   │   ├── chat/          # Chat interface components
│   │   ├── command-palette/ # Command palette
│   │   ├── layout/        # Main window layout
│   │   ├── preferences/   # Settings UI
│   │   └── ui/            # Base shadcn/ui components
│   ├── hooks/             # Custom React hooks
│   ├── lib/               # Utilities and services
│   ├── services/          # LLM and MCP services
│   ├── store/             # Zustand state stores
│   └── types/             # TypeScript definitions
├── src-tauri/             # Rust backend
│   ├── src/               # Tauri commands and logic
│   └── capabilities/      # Tauri permissions
└── docs/                  # Documentation

Tech Stack

• Frontend: React 19, TypeScript, Tailwind CSS v4, shadcn/ui v4
• State: Zustand v5, TanStack Query v5
• Backend: Tauri v2, Rust
• Build: Vite v7, Vitest v4

📋 TODO Before Release

• [ ] Generate app icons - Replace placeholder icons with branded assets
• [ ] Generate updater keys - Run tauri signer generate for auto-updates
• [ ] Code signing - Set up Apple Developer certificate for distribution
• [ ] Test on macOS - Verify on Intel and Apple Silicon

🤝 Contributing

Contributions are welcome! Please read [CONTRIBUTING.md](docs/CONTRIBUTING.md) for guidelines.

📄 License

This project is licensed under the [MIT License](LICENSE.md).

🙏 Acknowledgments

• Tauri - For the amazing desktop framework
• shadcn/ui - For beautiful UI components
• Model Context Protocol - For the MCP specification

---

Made with ❤️ by Navjot Dhanawat

Source & license

This open-source MCP server is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: navjotdhanawat
• Source: navjotdhanawat/nexus-ai
• License: MIT

Install and usage instructions live in the source repository linked above.