open-codex-computer-use: Automate Desktop Interfaces Through MCP

Summary
open-codex-computer-use exposes desktop automation as an MCP service that AI agents and compatible clients can use on macOS, Linux, and Windows. It is useful when an agent needs to interact with desktop apps through accessibility-based controls.
At a glance
- Language
- Swift
- License
- MIT
- Stars
- 2.3k
- Forks
- 204
- Added to OSRepos
- October 5, 2026
- Last analyzed
- October 5, 2026
Topics
Click on any tag to explore related repositories
Use at your own risk
OSRepos shares public repositories for knowledge and discovery only. Any installation, execution, configuration, or use of code from these repositories is the user's own responsibility. Always review the repository, source code, dependencies, licenses, and security implications before running or installing anything. OSRepos is not responsible for issues, damages, or losses resulting from third-party repositories.
Overview
open-codex-computer-use provides computer-use actions as an MCP service, letting compatible AI agents interact with desktop applications. It aims to make desktop automation available across macOS, Linux, and Windows without requiring a client-specific integration.
The project is a fit for developers who want to connect an agent such as Codex, Claude Code, Gemini CLI, or opencode to desktop workflows. Its interface includes MCP tools as well as CLI commands for individual calls and JavaScript-based automation.
Key Features
- Exposes computer-use capabilities through the Model Context Protocol (MCP).
- Supports desktop automation on macOS, Linux, and Windows.
- Provides installation commands for configuring several MCP clients.
- Includes CLI commands for calling tools, running sequences, and checking runtime capabilities.
- Offers
ocu jsand a terminal REPL for JavaScript automation withcuabindings. - Provides an installable skill for Codex and Claude Code.
Use Cases
- Agent developers can give Codex or another MCP-compatible client access to desktop applications.
- Automation builders can run repeatable sequences of desktop actions from the CLI.
- Teams working across operating systems can try a shared computer-use interface on macOS, Linux, and Windows.
- Developers prototyping agent workflows can use the JavaScript command or REPL to explore and test actions interactively.
Project Facts
- Language: Swift
- License: MIT
- Stars: 2.3k
- Forks: 204
- Topics: accessibility, ai-agent, ai-agents, claude-code, codex, codex-computer-use, computer-use, computer-use-agent, cross-platform, cua, desktop-automation, gemini-cli, gui-automation, macos, mcp, model-context-protocol, open-computer-use, open-computer-use-mcp, openai-codex, opencode
- Archived: No
Getting Started
Install the npm package:
npm i -g open-computer-use
Run open-computer-use or its ocu alias. On macOS, grant Accessibility and Screen Recording permissions. See the README for MCP client setup and further instructions.
Alternatives
- Open-Interface: Open-Interface uses screenshots and an LLM to guide simulated mouse and keyboard actions, while open-codex-computer-use exposes accessibility-based controls through MCP.
Considerations
- The macOS runtime requires macOS 14.0 or later and Accessibility and Screen Recording permissions.
- The npm launcher requires Node.js 18 or newer.
- While the project supports Windows and Linux as well as macOS, the README calls out additional permission setup specifically for macOS.
- The repository is MIT-licensed. Check the repository for current releases and platform-specific details.
Found this useful?
Share it with someone who would like open-codex-computer-use.
Source repository
Open the original repository on GitHub.
Related repositories
Similar repositories that may be relevant next.

agent-harness: Use Coding Agents from Rust Applications
October 5, 2026
A Rust framework that gives multiple coding agents a shared interface and typed event stream. Use it to embed agent runs in applications without writing separate output parsers for each provider or CLI.

image-blaster: Turn Images into 3D Environments and Assets
October 5, 2026
image-blaster uses Claude skills and external generation services to turn an image into a 3D environment, object meshes, and sound effects. It is designed to quickly create starting points for game, 3D, and web projects.

omnigent: Orchestrate AI Coding Agents Across Harnesses
October 5, 2026
Omnigent provides a shared orchestration layer for AI coding agents, with policies, sandboxing, and team collaboration. It suits developers who want to combine agent runtimes and access sessions across devices without tying workflows to one harness.

skillrank: Find and Evaluate AI-Agent Skills
October 4, 2026
SkillRank is a Rust CLI for discovering, installing, and evaluating skills used by coding agents. Its paired local evaluations compare a skill-enabled run with a control to show changes in success, tokens, time, and cost.