Qwen Desktop for Linux

Open-source Qwen AI (Tongyi Qianwen) desktop client for Linux. Run Alibaba's Qwen models natively on Ubuntu, Fedora, Arch, and all Linux distributions with full MCP (Model Context Protocol) support.

The official Qwen Desktop app only supports Windows and macOS. This project brings the same desktop experience to Linux.

Quick Install

Option 1: AppImage (recommended — works on every distro)

wget https://github.com/youssefvdel/qwen-desktop-linux/releases/latest/download/Qwen-1.1.2-x86_64.AppImage
chmod +x Qwen-1.1.2-x86_64.AppImage
./Qwen-1.1.2-x86_64.AppImage

Option 2: Debian/Ubuntu

wget https://github.com/youssefvdel/qwen-desktop-linux/releases/latest/download/qwen-desktop_1.1.2_amd64.deb
sudo apt install ./qwen-desktop_1.1.2_amd64.deb
qwen-desktop

Option 3: Fedora/RHEL

sudo dnf install https://github.com/youssefvdel/qwen-desktop-linux/releases/latest/download/qwen-desktop-1.1.2.x86_64.rpm
qwen-desktop

All Downloads

Features

Screenshots

Main chat interface with Skills menu open

MCP configuration panel

FAQ

What is Qwen AI?

Qwen (通义千问, Tongyi Qianwen) is a family of large language models developed by Alibaba Cloud's Tongyi Lab. It includes Qwen-Max, Qwen-Plus, Qwen3-6Plus, and Qwen3-235B-A22B.

Is this the official Qwen Desktop app?

No. The official app only supports Windows and macOS. This is a community-built, open-source desktop wrapper for Linux, reverse-engineered from the official app's protocol.

Is it free?

Yes. MIT license. No account needed beyond your Qwen account on chat.qwen.ai.

What distributions are supported?

All of them. AppImage works everywhere. We also ship native .deb and .rpm packages.

What is MCP?

Model Context Protocol lets Qwen interact with local tools and data. Enabled MCP servers can read/write files, automate your browser, query databases, and run CLI tools — all from chat.

Does it work on Wayland?

Yes, but some Wayland compositors need --enable-features=UseOzonePlatform --ozone-platform=wayland as a launch flag. X11 works out of the box.

Browser vs Desktop — why use this?

Architecture

Built with Electron 34 + TypeScript, mirroring the official Qwen Desktop app:

┌─────────────────────────────────────────────┐
│           Qwen Desktop (Electron)           │
│                                             │
│  ┌─────────────┐    ┌──────────────────┐    │
│  │ Main Process│◄──►│   MCP Proxy      │    │
│  │             │    │  (McpProxy)      │    │
│  │  - IPC      │    │  - stdio client  │    │
│  │  - Tray     │    │  - SSE client    │    │
│  │  - Menu     │    │  - HTTP stream   │    │
│  └──────┬──────┘    └────────┬─────────     │
│         │                    │              │
│  ┌──────▼──────┐    ┌────────▼─────────┐    │
│  │  Preload    │    │  Bundled Runtimes│    │
│  │  (Bridge)   │    │  - bun           │    │
│  │             │    │  - uv / uvx      │    │
│  └──────┬──────┘    └──────────────────┘    │
│         │                                   │
│  ┌──────▼───────────────────────────────┐   │
│  │         WebView                      │   │
│  │   chat.qwen.ai (desktop detection)   │   │
│  │   window.electronAPI exposed         │   │
│  └──────────────────────────────────────┘   │
└─────────────────────────────────────────────┘

Module Structure

Development

Prerequisites

• Node.js 22+
• npm
• Linux (x64 or ARM64)

Install & Run

git clone https://github.com/youssefvdel/qwen-desktop-linux.git
cd qwen-desktop-linux
npm install
npm start

Build Packages

# AppImage
npm run build

# All formats
npm run make

# Individual formats
npm run make:deb    # Debian/Ubuntu
npm run make:rpm    # Fedora/RHEL

Project Structure

qwen-desktop-linux/
├── src/
│   ├── main/           # Electron main process
│   │   ├── index.ts
│   │   ├── window-manager.ts
│   │   ├── ipc-handlers.ts
│   │   ├── skills-manager.ts
│   │   ├── app-lifecycle.ts
│   │   ├── runtime.ts
│   │   └── mcp-config.ts
│   ├── mcp/
│   │   ├── proxy.ts
│   │   └── server-client.ts
│   ├── preload/
│   │   └── index.ts
│   ├── renderer/
│   └── shared/
├── resources/
│   ├── bun/linux-x64/
│   └── uv/linux-x64/
├── dist/
└── package.json

MCP Configuration

Available MCP Servers

Example Config

Default MCP servers are created on first launch. Customize via the app's settings:

{
  "Filesystem": {
    "command": "bun",
    "args": [
      "x",
      "-y",
      "@modelcontextprotocol/server-filesystem",
      "/home/user/Documents"
    ],
    "transportType": "stdio"
  },
  "Fetch": {
    "command": "bun",
    "args": ["x", "-y", "@modelcontextprotocol/server-fetch"],
    "transportType": "stdio"
  },
  "Browser": {
    "command": "uvx",
    "args": ["@anthropic/mcp-server-playwright"],
    "transportType": "stdio"
  }
}

The app automatically replaces npx, bun, and uvx commands with bundled runtime paths.

Skills System

Create system prompts as .md files in ~/.config/qwen-desktop-linux/skills/. Inject into chat via the Skills menu.

Example

# Python Expert

You are a senior Python developer. Focus on:

- PEP8 compliance and type hints
- Performance optimization
- Error handling and testing
- Modern Python 3.12+ features

Menu → Skills → Python Expert injects the prompt.

Built-in Skills

• linux-power-user.md — Terminal commands and system administration
• code-reviewer.md — Code review with scoring
• python-expert.md — Python development guidance (sample)

Internationalization

zh-CN, en-US, zh-TW, ja-JP, ko-KR, ru-RU, de-DE, fr-FR, es-ES, it-IT, pt-PT, ar-BH

Security

Known Issues

Contributing

Contributions welcome.

1. Fork the repository
2. Create a feature branch (git checkout -b feature/your-feature)
3. Commit your changes
4. Push and open a Pull Request

What we need:

• Translations for additional languages
• Bug reports and fixes
• Documentation improvements
• UI/UX enhancements
• Testing on different distros

License

MIT License. See LICENSE.

Acknowledgments

• Based on reverse engineering of the official Qwen Desktop app (Windows/macOS)
• Uses @modelcontextprotocol/sdk by Anthropic
• Bundled runtimes: Bun + uv
• Built with Electron

Keywords: Qwen AI Linux, Qwen Desktop Linux, Tongyi Qianwen Linux, open source Qwen, free Qwen AI desktop, Alibaba Qwen Linux client, Qwen MCP Linux, Qwen3 Linux app, Qwen chat desktop Linux, Ubuntu Qwen AI, Fedora Qwen, Arch Linux Qwen, Electron Qwen app