Computer Use Linux is a local MCP server for AI agents that need to work in Linux desktop apps. It supports Wayland and uses AT-SPI accessibility data to find controls by their role, name or text, so agents can act on buttons and fields without relying entirely on screenshot coordinates. It's written in Rust and is open source under the MIT license.
Agents can inspect running apps, read window contents and capture screenshots of the desktop or a chosen window. They can also click, drag, scroll and type, focus specific windows, and change text fields or sliders through native accessibility actions. Coordinate input remains available for games, canvas surfaces and apps that don't expose accessible controls.
The desktop backends cover GNOME, KDE/KWin, Hyprland, niri, i3 and COSMIC, with best-effort X11 support. Actual capabilities depend on the control interfaces available in the desktop session. A structured diagnostics report identifies working backends and blockers, which helps distinguish a missing desktop capability from an agent error.
The server runs on your Linux machine and passes screenshots and accessibility data to the connected MCP host. It works with clients such as Hermes Agent and Claude Desktop, and has plugins for Claude Code and Codex. It originated in codex-desktop-linux, which also bundles it. A CLI exposes the same desktop capabilities for scripts; optional shell execution is disabled by default and runs with the server user's permissions when enabled.
Claim this page and we'll verify you by hand. Computer Use Linux gets the verified badge, and you can upgrade the listing to be featured on localhosted. Proud to be listed? Put our badge on your site.
Want more people to find Computer Use Linux?Promote it
Something wrong or outdated on this page?
27.6KUpdated 1 day agoMIT
macOS · Windows · Linux#Code execution#MCP#Tool calling
Cua gives AI agents access to computers they can inspect and operate, with tools for desktop automation, local virtual machines, and hosted fleets. It's for developers building agents that work across native apps and browsers, or evaluating how well those agents complete computer tasks. You bring the agent and model.
28.2KUpdated 1 day agoApache-2.0
Linux · Docker#Browser automation#Code execution#MCP
426Updated 3 weeks agoMIT
macOS · Windows · Linux · Docker · Web#Agent Skills#Code execution#Human approval
37.7KUpdated 3 days agoApache-2.0
macOS · Windows · Linux · Docker#Browser automation#Code execution#LM Studio integration
4.4KUpdated 1 day agoMIT
macOS · Windows · Linux · Web · JetBrains#Agent Client Protocol#Code execution#Git integration
6.2KUpdated 4 months agoMIT
macOS · Docker#Agent Skills#Browser automation#Code execution
Obscura is a headless browser engine for developers building AI agents and web scraping systems that need many independent browser sessions. Each session starts with fresh state, so cookies, logins, and cached data don't carry between tasks. It's written in Rust and runs without Chromium, with a smaller memory footprint intended to support more concurrent sessions per machine.
OwnPilot is a self-hosted personal AI assistant for people who want chat, personal records, and autonomous agents in one browser interface. It runs through Docker or on macOS, Windows, and Linux, and it's open source under the MIT license. Agents can work in separate workspaces, keep memory across tasks, and delegate work to other agents.
Playwright MCP lets AI agents control a browser by reading structured page information rather than interpreting screenshots. It's for developers building browser automation, exploratory tests or agent workflows that need to keep a browser session open across repeated actions. The server runs locally on macOS, Windows and Linux, or as a self-hosted service.
gptme is a self-hosted AI agent that works directly in your terminal, with access to your files and installed tools. It's for developers who want a coding assistant in their own environment, and people who want an agent for data analysis or other knowledge work. The software is free under the MIT license and doesn't require a gptme account.
bb-browser lets AI agents work through your local Chrome browser using the website accounts you're already signed into. It's for people building agents that need to research across websites or interact with authenticated pages without a separate API integration for each service. The project is open source under the MIT license.