UI TARS Desktop
📇 🏠 - Browser automation capabilities using Puppeteer, both support local and remote browser connection.
What is UI TARS Desktop?
UI TARS Desktop is a Model Context Protocol (MCP) server that allows AI assistants like Claude, Cursor, and VS Code to 📇 🏠 - browser automation capabilities using puppeteer, both support local and remote browser connection.
📇 🏠 - Browser automation capabilities using Puppeteer, both support local and remote browser connection.
This server falls under the Browser Automation category on MCPgee, the world's largest MCP server directory with 33,000+ servers.
Features
- 📇 🏠 - Browser automation capabilities using Puppeteer, both
Use Cases
Maintainer
Works with
Installation
NPM
npx -y @agent-infra/mcp-server-browserManual Installation
npx -y @agent-infra/mcp-server-browserConfiguration
Configuration Details
claude_desktop_config.json
Performance
Response Metrics
Resource Usage
How to Set Up and Use UI TARS Desktop
The UI-TARS Desktop MCP browser server (@agent-infra/mcp-server-browser) provides AI assistants with full Puppeteer-based browser automation capabilities, supporting both local and remote browser connections. It is part of ByteDance's Agent TARS multimodal agent stack and exposes tools for navigation, clicking, typing, screenshotting, and JavaScript evaluation through the MCP protocol. Developers use it to give Claude and other MCP clients the ability to control a real browser for web scraping, automated testing, and GUI-driven workflows.
Prerequisites
- Node.js 18 or later with npx available
- Google Chrome or Chromium installed (Puppeteer will attempt to use system Chrome)
- An MCP-compatible client such as Claude Desktop, Cursor, or Agent TARS
- Optional: a running remote Chrome instance with remote debugging enabled for remote browser connections
Install and run the MCP browser server
The server is published as @agent-infra/mcp-server-browser on npm. Run it directly with npx — it will launch a Puppeteer-controlled browser on startup.
npx -y @agent-infra/mcp-server-browserAdd the server to your MCP client configuration
Edit your claude_desktop_config.json (Claude Desktop) or equivalent config file to register the browser server.
{
"mcpServers": {
"browser": {
"command": "npx",
"args": ["-y", "@agent-infra/mcp-server-browser"]
}
}
}Configure a remote browser connection (optional)
To connect to an existing Chrome instance instead of launching a new one, start Chrome with remote debugging enabled and pass the WebSocket endpoint to the server.
# Launch Chrome with remote debugging
chrome --remote-debugging-port=9222 --user-data-dir=/tmp/chrome-debug
# Then configure the MCP server to connect remotely
{
"mcpServers": {
"browser": {
"command": "npx",
"args": ["-y", "@agent-infra/mcp-server-browser"],
"env": {
"BROWSER_WS_ENDPOINT": "ws://localhost:9222"
}
}
}
}Restart your MCP client
Save the config file and restart Claude Desktop or your MCP client. The browser tools should appear in the available tools list.
Test the browser connection
Ask your AI assistant to navigate to a URL. The Puppeteer browser should open and load the page, with the result returned to the AI client.
UI TARS Desktop Examples
Client configuration
Minimal claude_desktop_config.json entry for the UI-TARS browser MCP server using local Puppeteer.
{
"mcpServers": {
"browser": {
"command": "npx",
"args": ["-y", "@agent-infra/mcp-server-browser"]
}
}
}Prompts to try
Example prompts that leverage the Puppeteer browser automation tools.
- "Navigate to https://news.ycombinator.com and list the top 5 headlines"
- "Take a screenshot of https://example.com"
- "Go to https://github.com/trending and scrape the top 10 trending repositories"
- "Fill in the login form at https://app.example.com with username 'test' and submit"
- "Evaluate document.title on the current page"Troubleshooting UI TARS Desktop
Puppeteer fails to launch with 'Cannot find Chrome executable'
Install Chrome or Chromium on your system. Alternatively, let Puppeteer download its own Chromium by running 'npx puppeteer browsers install chrome' before starting the server.
Remote browser connection times out with BROWSER_WS_ENDPOINT set
Verify Chrome is running with '--remote-debugging-port=9222' and check the WebSocket URL by visiting http://localhost:9222/json/version in a browser. Ensure no firewall blocks the port.
Server starts but no browser tools appear in the MCP client
Ensure the MCP client was fully restarted after editing the config file. Check the server process is running with 'ps aux | grep mcp-server-browser'. Look at the MCP client's logs for connection errors.
Frequently Asked Questions about UI TARS Desktop
What is UI TARS Desktop?
UI TARS Desktop is a Model Context Protocol (MCP) server that 📇 🏠 - browser automation capabilities using puppeteer, both support local and remote browser connection. It connects AI assistants to external tools and data sources through a standardized interface.
How do I install UI TARS Desktop?
Install via npm with the command: npx -y @agent-infra/mcp-server-browser. Then add the server configuration to your AI client's JSON config file (e.g., claude_desktop_config.json or .cursor/mcp.json).
Which AI clients work with UI TARS Desktop?
UI TARS Desktop works with all major MCP-compatible AI clients including Claude Desktop, Claude Code, Cursor, VS Code (GitHub Copilot), Windsurf, and Cline.
Is UI TARS Desktop free to use?
Yes, UI TARS Desktop is open source and available under the Apache-2.0 license. You can use it freely in both personal and commercial projects.
UI TARS Desktop Alternatives — Similar Browser Automation Servers
Looking for alternatives to UI TARS Desktop? Here are other popular browser automation servers you can use with Claude, Cursor, and VS Code.
Chrome DevTools MCP
★ 40.6kAI-powered Chrome automation server with natural language element detection. Control Chrome browser through MCP protocol for testing, debugging, and performance analysis. Features 91% accuracy in element location, works with free AI models, and suppo
Playwright
★ 32.8kA production-ready browser automation server that enables AI assistants to interact with web pages using tools for navigation, element interaction, and data extraction. It features a built-in Inspector UI and robust crash recovery for reliable automa
Page Agent
★ 18.0kJavaScript in-page GUI agent. Control web interfaces with natural language.
Chrome
★ 11.7kAn extension-based MCP server that enables AI assistants to control your browser, leveraging existing sessions and login states for automation and content analysis. It provides over 20 tools for semantic tab search, interactive element manipulation,
LAMDA
★ 7.8kThe most powerful Android RPA agent framework, next generation mobile automation.
Browser Tools MCP
★ 7.2kThis application is a powerful browser monitoring and interaction tool that enables AI-powered applications via Anthropic's Model Context Protocol (MCP) to capture and analyze browser data through a Chrome extension.
Browse More Browser Automation MCP Servers
Explore all browser automation servers available in the MCPgee directory. Each server includes setup guides for Claude, Cursor, and VS Code.
Set Up UI TARS Desktop in Your Editor
Choose your AI client for step-by-step setup instructions.
Quick Config Preview
Add this to your claude_desktop_config.json or .cursor/mcp.json
Ready to use UI TARS Desktop?
Browse our complete directory of 33,000+ MCP servers, read setup guides for your editor, and start building with the Model Context Protocol.