UI-TARS-desktop
The Open-Source Multimodal AI Agent Stack: Connecting Cutting-Edge AI Models and Agent Infra
npx -y monorepo{
"mcpServers": {
"ui-tars-desktop": {
"command": "npx",
"args": [
"-y",
"monorepo"
]
}
}
}UI-TARS-desktop is an officially maintained MCP server in the AI & ML category, developed by bytedance. It runs locally on your machine, keeping your data private and giving you full control over the connection. AI engineers can use it to chain models and pipelines into more powerful workflows.
About UI-TARS-desktop
Overview
The Open-Source Multimodal AI Agent Stack: Connecting Cutting-Edge AI Models and Agent Infra
Links
Topics
agent, agent-tars, browser-use, computer-use, cowork, gui-agent, gui-operator, mcp, mcp-server, multimodal, tars, ui-tars, vision, vlm
Who Should Use UI-TARS-desktop?
- 1Control a browser and scrape the web through Claude
- 2Chain AI models and pipelines through a unified MCP interface
- 3Let Claude orchestrate other AI tools and models
- 4Integrate embeddings, image generation, or speech APIs into your workflow
How to Install UI-TARS-desktop
Before you start
You will need Node.js (v18 or later) installed on your machine — download it from nodejs.org if you haven't already.
- 1Open a terminal (Terminal on Mac, Command Prompt or PowerShell on Windows).
- 2Paste the install command above and press Enter — Node.js will download and run the server automatically.
- 3Add the server to your Claude Desktop config file (see the JSON snippet above) and restart Claude.
The Claude Desktop config snippet above can be copied and pasted directly into your claude_desktop_config.json file — no editing required.