What is CUA?

CUA (Computer-Use Agent) is a MCP server that allows users to control their computer through natural language commands using MCP clients like Claude Desktop or Cursor. It bridges human instructions and computer actions, enabling automated computer control.

How to use CUA?

1. Prerequisites: Set up full CUA capabilities by installing the Lume CLI, pulling the latest macOS CUA image, and starting the Lume daemon service.
2. Installation: Install via PyPI (

pip install cua-mcp-server
```) or use the easy setup script (
```bash
curl -fsSL https://raw.githubusercontent.com/trycua/cua/main/libs/mcp-server/scripts/install_mcp_server.sh | bash
```).
3. **Configuration**: Add the CUA agent to your MCP client's configuration (e.g., Claude Desktop or Cursor) with the required environment variables.
4. **Usage**: Issue natural language commands to the MCP client, such as "Open Chrome and go to github.com."

## Key features of CUA
- **Computer control via natural language**: Execute tasks like opening apps, managing files, or capturing screenshots using text commands.
- **MCP client integration**: Works with Claude Desktop, Cursor, or other MCP-compatible clients.
- **Flexible configuration**: Customize settings like the agent loop, model provider, and API keys.
- **Development-friendly**: Supports local development and testing without installation.

## Use cases of CUA
1. **File management**: Automate tasks like creating folders, searching for files, or organizing downloads.
2. **App control**: Open or close applications, navigate websites, or interact with software.
3. **Screen capture**: Take and annotate screenshots or record screen activity.
4. **Workflow automation**: Define multi-step processes (e.g., opening an app, navigating to a URL, and logging in).

## FAQ from CUA
- **What MCP clients are compatible with CUA?**
> CUA supports Claude Desktop, Cursor, and other MCP-compatible clients.

- **How do I troubleshoot installation errors?**
> Use absolute paths for scripts and check logs with 
```bash
tail -n 20 -f ~/Library/Logs/Claude/mcp*.log
```.

- **Can I use CUA without API keys?**
> No, you need valid API keys for the chosen model provider (e.g., Anthropic or OpenAI).

cua-mcp-server is a MCP server for the Computer-Use Agent (CUA), allowing you to run CUA through Claude Desktop or other MCP clients.

Get started with Agent

Prerequisites

Before installing the MCP server, you'll need to set up the full Computer-Use Agent capabilities as described in Option 2 of the main README. This includes:

1. Installing the Lume CLI
2. Pulling the latest macOS CUA image
3. Starting the Lume daemon service
4. Installing the required Python libraries (Optional: only needed if you want to verify the agent is working before installing MCP server)

Make sure these steps are completed and working before proceeding with the MCP server installation.

Installation

Install the package from PyPI:

pip install cua-mcp-server

This will install:

• The MCP server
• CUA agent and computer dependencies
• An executable cua-mcp-server script in your PATH

Easy Setup Script

If you want to simplify installation, you can use this one-liner to download and run the installation script:

curl -fsSL https://raw.githubusercontent.com/trycua/cua/main/libs/mcp-server/scripts/install_mcp_server.sh | bash

This script will:

• Create the ~/.cua directory if it doesn't exist
• Generate a startup script at ~/.cua/start_mcp_server.sh
• Make the script executable
• The startup script automatically manages Python virtual environments and installs/updates the cua-mcp-server package

You can then use the script in your MCP configuration like this:

{ 
  "mcpServers": {
    "cua-agent": {
      "command": "/bin/bash",
      "args": ["~/.cua/start_mcp_server.sh"],
      "env": {
        "CUA_AGENT_LOOP": "OMNI",
        "CUA_MODEL_PROVIDER": "ANTHROPIC",
        "CUA_MODEL_NAME": "claude-3-7-sonnet-20250219",
        "CUA_PROVIDER_API_KEY": "your-api-key"
      }
    }
  }
}

Development Guide

If you want to develop with the cua-mcp-server directly without installation, you can use this configuration:

{
  "mcpServers": {
    "cua-agent": {
      "command": "/bin/bash",
      "args": ["~/cua/libs/mcp-server/scripts/start_mcp_server.sh"],
      "env": {
        "CUA_AGENT_LOOP": "UITARS",
        "CUA_MODEL_PROVIDER": "OAICOMPAT",
        "CUA_MODEL_NAME": "ByteDance-Seed/UI-TARS-1.5-7B",
        "CUA_PROVIDER_BASE_URL": "https://****************.us-east-1.aws.endpoints.huggingface.cloud/v1",
        "CUA_PROVIDER_API_KEY": "your-api-key"
      }
    }
  }
}

This configuration:

• Uses the start_mcp_server.sh script which automatically sets up the Python path and runs the server module
• Works with Claude Desktop, Cursor, or any other MCP client
• Automatically uses your development code without requiring installation

Just add this to your MCP client's configuration and it will use your local development version of the server.

Troubleshooting

If you get a /bin/bash: ~/cua/libs/mcp-server/scripts/start_mcp_server.sh: No such file or directory error, try changing the path to the script to be absolute instead of relative.

To see the logs:

tail -n 20 -f ~/Library/Logs/Claude/mcp*.log

Claude Desktop Integration

To use with Claude Desktop, add an entry to your Claude Desktop configuration (claude_desktop_config.json, typically found in ~/.config/claude-desktop/):

For more information on MCP with Claude Desktop, see the official MCP User Guide.

Cursor Integration

To use with Cursor, add an MCP configuration file in one of these locations:

• Project-specific: Create .cursor/mcp.json in your project directory
• Global: Create ~/.cursor/mcp.json in your home directory

After configuration, you can simply tell Cursor's Agent to perform computer tasks by explicitly mentioning the CUA agent, such as "Use the computer control tools to open Safari."

For more information on MCP with Cursor, see the official Cursor MCP documentation.

First-time Usage Notes

API Keys: Ensure you have valid API keys:

• Add your Anthropic API key, or other model provider API key in the Claude Desktop config (as shown above)
• Or set it as an environment variable in your shell profile

Configuration

The server is configured using environment variables (can be set in the Claude Desktop config):

Variable	Description	Default
CUA_AGENT_LOOP	Agent loop to use (OPENAI, ANTHROPIC, UITARS, OMNI)	OMNI
CUA_MODEL_PROVIDER	Model provider (ANTHROPIC, OPENAI, OLLAMA, OAICOMPAT)	ANTHROPIC
CUA_MODEL_NAME	Model name to use	None (provider default)
CUA_PROVIDER_BASE_URL	Base URL for provider API	None
CUA_MAX_IMAGES	Maximum number of images to keep in context	3

Available Tools

The MCP server exposes the following tools to Claude:

1. run_cua_task - Run a single Computer-Use Agent task with the given instruction
2. run_multi_cua_tasks - Run multiple tasks in sequence

Usage

Once configured, you can simply ask Claude to perform computer tasks:

• "Open Chrome and go to github.com"
• "Create a folder called 'Projects' on my desktop"
• "Find all PDFs in my Downloads folder"
• "Take a screenshot and highlight the error message"

Claude will automatically use your CUA agent to perform these tasks.