FREEDOM251

Freedom251 › AI Computer Use Agents

open-computer-use

Free plan APILinuxMacSelf-hostedWindows

Visit open-computer-use Alternatives Pricing

open-computer-use's own site

What open-computer-use is

Open Computer Use is a local MCP server for AI agents to inspect and operate desktop applications through accessibility interfaces. Its nine core tools cover discovering apps, capturing accessibility trees and screenshots, clicking, scrolling, dragging, typing, pressing keys, secondary actions and setting control values. The project says it runs entirely on the user’s machine and uses accessibility interfaces for core app actions without taking over the real mouse and keyboard. Setup instructions cover Codex CLI and App, Claude Code, Gemini CLI and opencode; other MCP clients can be configured manually. The repository also includes a ZCode plugin to install the skill and connect an MCP server. It is free and distributed under the MIT license, with support for macOS, Windows and Linux. macOS 14 or later requires Accessibility and Screen Recording permissions. Windows and Linux use a signed-in desktop session, and Linux requires AT-SPI2. Additional window-level tools are currently available on Windows, with macOS and Linux support in progress. Password managers are refused; app launching, focus changes and global input injection require explicit environment-variable gates.

Who it suits

Open Computer Use is for users building AI-agent workflows that interact with desktop apps through accessibility interfaces. It is suited to users able to configure local software and required system permissions.

What it does well

  • Runs locally on the user's machine.
  • Core tools cover app discovery and desktop actions.
  • Setup guidance names four MCP clients.
  • Free software under the MIT license.

What to know first

  • macOS 14 or later requires two permissions.
  • Linux requires AT-SPI2 and a signed-in desktop session.
  • Additional window-level tools are currently Windows-only.

Limits

Additional window-level tools are currently available on Windows; macOS and Linux support is in progress. macOS 14 or later requires Accessibility and Screen Recording permissions, and Linux requires AT-SPI2 in a signed-in desktop session.

Plans

PlanBilled and limitsPrice
Open sourceMIT-licensed · runs locally · macOS, Windows, LinuxFree

As github.com lists them, read 9 Oct 2026. The maker's own page has the final word.

Getting started

  1. Open the project repository on GitHub.
  2. Choose a documented setup command for Codex CLI and App, Claude Code, Gemini CLI, or opencode, or configure another MCP client manually.
  3. On macOS 14 or later, grant Accessibility and Screen Recording permissions.
  4. On Windows or Linux, use a signed-in desktop session; on Linux, ensure AT-SPI2 is available.
  5. Use the ZCode plugin if you want it to install the skill and connect an MCP server.

Verdict

Open Computer Use provides a broad set of local desktop controls for AI agents. Review platform permissions and the current Windows-only window-level tools before adopting it.

Questions

Is Open Computer Use free?

Yes. It is open source under the MIT license, and the listed plan is free.

Which operating systems does it support?

It supports macOS, Windows, and Linux. Additional window-level tools are currently available on Windows, with support for macOS and Linux in progress.

Which AI clients can connect to it?

The project documents setup for Codex CLI and App, Claude Code, Gemini CLI, and opencode. Other MCP clients can be configured manually.

What permissions does it need?

On macOS 14 or later, it requires Accessibility and Screen Recording permissions. Windows and Linux require a signed-in desktop session, and Linux also requires AT-SPI2.

Does it run locally?

Yes. The project says it runs entirely on the user’s machine and performs its core app actions through the accessibility layer.

What safety controls does it include?

It refuses password managers. App launching, focus changes, and global input injection require explicit environment-variable gates.

Facts from open-computer-use's own pages

Desktop actions
Its nine core tools support app discovery, accessibility-tree and screenshot capture, clicking, scrolling, dragging, typing, key presses, secondary actions, and setting control values. · github.com, 9 Oct 2026
Integrations
The project documents setup commands for Codex CLI and App, Claude Code, Gemini CLI, and opencode, and supports manual configuration for other MCP clients. · github.com, 9 Oct 2026
License
The repository is distributed under the MIT license. · github.com, 9 Oct 2026
Local operation
The project says it runs entirely on the user's machine and uses the accessibility layer without taking over the real mouse and keyboard for its core app actions. · github.com, 9 Oct 2026
Permissions
On macOS 14 or later, the project requires Accessibility and Screen Recording permissions; Windows and Linux work in a signed-in desktop session, with Linux requiring AT-SPI2. · github.com, 9 Oct 2026
Platform limitation
The additional window-level tools are currently available on Windows, with macOS and Linux support in progress. · github.com, 9 Oct 2026
Platforms
The runtime supports macOS, Windows, and Linux, with nine core tools available across all three. · github.com, 9 Oct 2026
Purpose
Open Computer Use is a local MCP server that lets AI agents inspect and operate desktop apps through accessibility interfaces. · github.com, 9 Oct 2026
Safety controls
Password managers are refused, while app launching, focus changes, and global input injection require explicit environment-variable gates. · github.com, 9 Oct 2026
Security policy
The security policy says fixes are provided for the latest npm release and asks users to report suspected vulnerabilities through GitHub's private vulnerability reporting channel. · github.com, 9 Oct 2026
Security scope
The policy notes that the tool performs high-privilege accessibility and input automation in the local signed-in desktop session and should only be used on authorized machines. · github.com, 9 Oct 2026
Support and audience
The project provides installation and troubleshooting references and invites contributions; its security policy says fixes are delivered through the regular release process. · github.com, 9 Oct 2026
ZCode plugin
The repository includes a ZCode plugin that installs the skill and connects an MCP server. · github.com, 9 Oct 2026

Where open-computer-use stands on Freedom251

Look at these too