Claude Computer Use is now in Claude Code: Everything You Need to Know (2026)

Anthropic just launched one of its most-requested features: Claude can now open your apps, click through your UI, and test what it built - all from the command line. Here's everything you need to know.

What Is Claude Computer Use?

Claude computer use is the ability for Claude to directly control a computer's graphical interface — clicking buttons, typing in fields, opening applications, navigating menus, and reading the screen — all autonomously, using instructions you give in plain English.

The capability was originally introduced in October 2024, when Anthropic released a "computer use" feature that allowed Claude to navigate computers by interpreting screen content and simulating keyboard and mouse input. It was initially a developer-only API beta. Now, in 2026, it has been folded directly into Claude Code — the CLI tool millions of developers already use every day.

Claude can open your apps, click through your UI, and test what it built, right from the CLI.

— Anthropic, March 2026

This is a pivotal shift. Computer use is no longer something you need to wire up via the raw API with a Docker container and custom screenshot pipelines. It's now a first-class citizen inside the tool you already have installed.

How to Enable Claude Computer Use (Step-by-Step)

Image- Claude

Getting started with the Claude computer use tool takes under two minutes if you're already a Claude Pro or Max subscriber. Here's the exact process:

1. Open the Claude Desktop App The computer use feature lives in the Desktop app (Cowork or Claude Code tab), not just the browser. Make sure you have the latest version installed from claude.ai.
2. Navigate to Settings → Desktop app → General If you ask Claude to do something that needs computer use while it's off, Claude tells you it could do the task if you enable it in Settings. To enable it, open Settings → Desktop app → General and toggle Computer use on.
3. Grant macOS System Permissions Before the toggle takes effect, you need to grant two macOS system permissions: Accessibility (lets Claude click, type, and scroll) and Screen Recording (lets Claude see what's on your screen). The Settings page shows the current status of each permission.
4. Start a Claude Code or Cowork Session Open Cowork or Claude Code in the desktop app and start a session. Ask Claude to do something that involves an app on your computer. Claude will ask for permission to access the app before proceeding.
5. Confirm and Monitor Claude checks each action before executing. Watch the session log in real time. You can interrupt at any point.

Claude Computer Use in Claude Code CLI

The Claude Code CLI computer use integration is the headline feature here. Developers who prefer to live in the terminal now get the full loop: write code, run tests, and visually QA the result — without ever switching out of their workflow.

A typical session might look like this:

# Inside a Claude Code session
> Build a Electron desktop app that shows local weather
# Claude writes the code, installs deps, builds the app

> Now open the app and verify the UI looks correct
# Claude uses computer use to launch the app window,
# take a visual screenshot, and report back

This closes the development loop that's long been broken in AI coding tools: you could generate code, but verifying it visually still required a human. Claude Code CLI computer use solves that.

How Claude Decides When to Use Computer Use

Claude has several ways to interact with an app or service, and computer use is the broadest and slowest. Screen control is reserved for things nothing else can reach — like native apps, hardware control panels, the iOS simulator, or proprietary tools without an API.

The priority order Claude follows is:

1. MCP Connectors — fastest and most reliable (Gmail, Slack, GitHub, etc.)
2. Browser via Claude in Chrome — used when there's no connector for the target site
3. Computer Use (Screen Control) — the fallback for native desktop apps and anything without an API

Computer Use Is Now in Claude Code: MCP Integration

One of the most powerful aspects of this launch is how Claude computer use integrates with MCP (Model Context Protocol). The combination unlocks genuinely agentic workflows.

In Cowork, Claude uses the most precise tool first. When you assign a task, Claude follows this order: Connectors (if a connector is available — like Gmail, Google Drive, or Slack — Claude uses it, which is the fastest and most reliable path); then Browser (when there isn't a connector for the tool you need, Claude can navigate the Chrome browser); and finally Screen interaction (Claude uses computer use to interact directly with your screen).

The practical result: you can build automated workflows like "check my inbox for the design brief, open Figma, screenshot the latest mockup, then write a review in Notion" — and Claude handles each step using the most efficient method available.

Claude Computer Use on Windows: What You Need to Know

If you're a Windows user asking whether Claude computer use works on your machine: the short answer is not yet, but it's coming.

Computer use is in research preview for Pro and Max plans. It's available in Cowork and Claude Code in the macOS Claude Desktop application. Windows support is coming soon.

In the meantime, Windows users can still access the full Claude Code CLI with all its coding capabilities. Since 2025, Claude Code runs natively on Windows 10 and Windows 11 without WSL. You only need Git for Windows installed. The computer use layer will be added to the Windows desktop app in a future update.

Claude Code on Windows: Quick Install

# Windows PowerShell (run as Administrator)
# Step 1: Install Git for Windows first (git-scm.com)

# Step 2: Install Claude Code natively
irm https://claude.ai/install.ps1 | iex

# Step 3: Verify installation
claude --version

# Step 4: Start a session
claude

Plan Availability: Who Gets Claude Computer Use?

Users on Pro and Max plans can give Claude access to computer use. Claude can open files, run dev tools, point, click, and navigate to what's on your screen to perform tasks itself — with no setup required.

Safety & What Claude Won't Touch

Computer use is powerful, which makes safety non-negotiable. Anthropic has built several guardrails into the research preview:

Unlike the sandboxed Bash tool, computer use runs on your actual desktop with access to whatever you approve. Claude checks each action and flags potential prompt injection from on-screen content, but the trust boundary is different.

We don't recommend using it with apps containing sensitive data relating to your healthcare, finances, or other personal records. Your desktop must be active — your computer needs to be awake and the Claude Desktop app needs to be open for computer use to work.

Use Cases: What You Can Actually Do Right Now

Here are practical, real-world scenarios where Claude computer use inside Claude Code delivers immediate value:

1. End-to-End UI Testing

Build a web app or native app with Claude Code, then have Claude open a browser or simulator, navigate the flows, and write a QA report — without you leaving the terminal.

2. iOS Simulator Testing

Computer use lets Claude open your apps, control your screen, and work directly on your machine the way you would. Ask Claude to test a native app in the iOS simulator, interact with a desktop tool that has no CLI, or automate something that only works through a GUI.

3. Desktop Tool Automation

Automate repetitive tasks in proprietary software that doesn't have an API or MCP connector — design tools, data entry systems, legacy internal apps.

4. Screenshot-Driven Iteration

Claude can take a screenshot, analyze the visual output, identify bugs or design issues, make code changes, and re-check — all in a single agent loop.

5. Dispatch + Computer Use

Claude's new computer use capability makes Dispatch even more helpful by allowing it to use your computer on your behalf while you're away. Schedule a long-running task, step away, and come back to a completed result.

What's New in the Claude Computer Use Tool (API Updates)

For developers building on top of the computer use API, the tool has also been updated. The computer_20250124 version introduced updated computer use tool commands including hold_key, left_mouse_down, left_mouse_up, scroll, triple_click, and wait.

The latest models to support computer use include Claude Opus 4.6, Sonnet 4.6, and Haiku 4.5 — with the beta header computer-use-2025-11-24 required for the most recent tool version.

curl https://api.anthropic.com/v1/messages \
  -H "anthropic-beta: computer-use-2025-11-24" \
  -d '{
    "model": "claude-sonnet-4-6",
    "tools": [
      {
        "type": "computer_20251124",
        "name": "computer",
        "display_width_px": 1920,
        "display_height_px": 1080
      }
    ]
  }'

FAQ: Claude Computer Use

Conclusion

Claude computer use arriving in Claude Code is the most significant developer workflow upgrade Anthropic has launched in 2026. For the first time, a single CLI tool can write code, run it, open the resulting app, visually verify it, catch bugs, fix them, and re-check — with no human in the loop between those steps.

It's a research preview, which means rough edges are expected. But the core capability is real, the safety architecture is thoughtful, and the MCP integration means the surface area of what Claude can automate grows with every new connector added to the ecosystem.

If you're on a Pro or Max plan and you haven't enabled it yet — you're leaving a significant chunk of productivity on the table.