Open-source infrastructure for Computer-Use Agents. Sandboxes, SDKs, and benchmarks to train and evaluate AI agents that can control full desktops (macOS, Linux, Windows).

Bytebot is a self-hosted AI desktop agent that automates computer tasks through natural language commands, operating within a containerized Linux desktop environment.

A research prototype of a human-centered web agent

Agent S: an open agentic framework that uses computers like a human

The first open-source Artificial Narrow Intelligence generalist agentic framework Computer-Using-Agent that fully operates graphical-user-interfaces (GUIs) by using only natural language. Uses Visualization-of-Thought and Chain-of-Thought reasoning to elicit spatial reasoning and perception, emulates, plans and simulates synthetic HID interactions.

Surf is a computer use AI agent powered by OpenAI that interacts with a E2B's virtual desktop environment through natural language instructions

[NeurIPS 2025] Mind2Web-2 Benchmark: Evaluating Agentic Search with Agent-as-a-Judge

This is the official website for TuriX Computer-use-Agent

ncurses (pdcurses) based library for CUA look menu

Emacs mode that moves to modern keybindings

A lightweight Python library to create, run and manage macOS and Linux virtual machines (VMs) natively on Apple Silicon.

Mark web pages for use with vision-language models

Opensource benchmark evaluating web operators/agents performance

A command line text editor with notepad like key bindings.

OS-Harm: A Benchmark for Measuring Safety of Computer Use Agents [NeurIPS 2025 Spotlight]

Fully self hosted chatgpt agent alternative

Browser using agent (bua) notte model on any browser environment

Emulate standard keybindings from modern editors

A simple terminal source code editor.

CoAct-1: Computer-using Agents with Coding as Actions