Quickstart

This guide gets you from zero to a working computer-use agent using Cua Cloud sandboxes.

Prerequisites: Python 3.12 or 3.13 and a free Cua account at cua.ai.

1. Install the Cua CLI and SDKs

pip install cua-cli cua-computer "cua-agent[all]"

2. Authenticate

cua auth login

This opens your browser for authentication. Once complete, your API key is stored locally.

To export your API key to a .env file for use in scripts:

cua auth env

3. Create a Cloud Sandbox

Create a Linux sandbox:

cua sandbox create --os linux --size small --region north-america

Note the sandbox name from the output (e.g., curious-fox-123).

You can view all your sandboxes:

cua sandbox list

4. Take a Screenshot (Hello World)

Create a file hello.py:

import os
import asyncio
from computer import Computer

# Load API key from environment (or set directly)
os.environ["CUA_API_KEY"] = "sk_cua-api01_..."  # or use: cua auth env

computer = Computer(
    os_type="linux",
    provider_type="cloud",
    name="curious-fox-123"  # your sandbox name
)

async def main():
    await computer.run()  # Connect to the sandbox

    try:
        # Take a screenshot
        screenshot = await computer.interface.screenshot()
        screenshot.save("screenshot.png")
        print("Screenshot saved to screenshot.png")

        # Click at coordinates (100, 100)
        await computer.interface.left_click(100, 100)

        # Type some text
        await computer.interface.type_text("Hello from Cua!")
    finally:
        await computer.disconnect()

asyncio.run(main())

Run it:

python hello.py

5. Run an Agent

Now let's add AI. Create agent.py:

import os
import asyncio
from computer import Computer
from agent import ComputerAgent

os.environ["CUA_API_KEY"] = "sk_cua-api01_..."  # or use: cua auth env

computer = Computer(
    os_type="linux",
    provider_type="cloud",
    name="curious-fox-123"  # your sandbox name
)

async def main():
    await computer.run()

    try:
        agent = ComputerAgent(
            model="cua/anthropic/claude-sonnet-4.5",
            tools=[computer],
        )

        messages = [{"role": "user", "content": "Open Firefox and go to google.com"}]

        async for result in agent.run(messages):
            for item in result["output"]:
                if item["type"] == "message":
                    print(item["content"][0]["text"])
    finally:
        await computer.disconnect()

asyncio.run(main())

Run it:

python agent.py

The agent will observe the screen, decide what actions to take, and execute them to complete your task.

6. Open the Desktop in Your Browser

Watch your agent work in real-time:

cua sandbox vnc curious-fox-123

This opens a browser window showing the sandbox desktop.

7. Clean Up

When you're done, stop your sandbox to avoid charges:

cua sandbox stop curious-fox-123

Or delete it entirely:

cua sandbox delete curious-fox-123

CLI Reference

Command	Description
`cua auth login`	Authenticate with your Cua account
`cua auth env`	Export API key to `.env` file
`cua sandbox list`	List all sandboxes
`cua sandbox create --os <os> --size <size> --region <region>`	Create a sandbox
`cua sandbox vnc <name>`	Open sandbox in browser
`cua sandbox start <name>`	Start a stopped sandbox
`cua sandbox stop <name>`	Stop a sandbox
`cua sandbox delete <name>`	Delete a sandbox

Next Steps

Using the Computer SDK - Learn low-level computer control
Using the Agent SDK - Advanced AI agent configuration
Self-Hosted Sandboxes - Run sandboxes locally with Docker

Was this page helpful?

Quickstart

On this page