computer-use-agents

Mano-P: Open-source GUI-VLA agent for edge devices. #1 on OSWorld (specialized, 58.2%). Runs locally on Apple M4 Mac mini/MacBook — no data leaves your device.Mano-P 是一个开源 GUI-VLA 项目，支持在 Mac mini/MacBook 上或通过算力棒本地运行推理，实现纯视觉驱动的跨平台 GUI 自动化操作。数据完全本地处理，支持复杂多步骤任务规划与执行。

Agent Desktop gives any agent reliable computer use on the desktop. Built with Rust, it sees any app's real UI structure through OS accessibility trees and operates it — refs stay stable and actions stay safe to retry, instead of guessing from pixels.

For EHRs, payer portals, and any app your team clicks through. A person or an agent shows the task; OpenAdapt turns it into a local program that repeats the steps with no model calls, checks each save, and stops when something doesn't match.

[ICLR 2026 Oral] ScaleCUA is the open-sourced computer use agents that can operate on cross-platform environments (Windows, macOS, Ubuntu, Android).

"AnyTool: Universal Tool-Use Layer for AI Agents"

Automatic computer use agent for end-to-end automation workflows on macOS, Windows, Linux, iOS and the browser.

Open-source alternative to Claude Co-Work built by Claude. with BYOK

[ECCV 2026] Official Repository for "DualTAP: A Dual-Task Adversarial Protector for Mobile MLLM Agents"

A Swift library for macOS automation — mouse, keyboard, screenshots, image recognition, and AI-powered agents.

Official code for "Relevance Does Not Imply Applicability: Experience Activation for Personal GUI Agents".

Open-source framework for computer use agents: VeriGen verifiable task synthesis, online RL training (AgentRL), and OSWorld/ScienceBoard evaluation.

let LLMs control your desktop for tasks

Open, source-backed catalog of computer-use agents: reviewed papers, GUI models, datasets, benchmarks, tools, and reusable JSONL/SQLite data.

Repository for the implementation of MacArena

AOI: a model-agnostic perception layer giving computer-use agents adaptive video + audio perception with zero retraining, plus DynaCU-Bench (100 dynamic browser tasks). Paper + benchmark + code.

DeskForge: Dense Supervision from Desktop Environments for Computer-Use Agents

Papers, benchmarks, demos, and tools for multimodal coding agents: visual feedback for UI, CAD, 3D, games, charts, and robotics.

Computer Use Examples with API Integration Full Stack

A Lightweight Desktop GUI Agent via Dynamic Focus Vision and Hierarchical Memory. Your AI-powered hands and eyes for desktop automation.

A clicker for LLMs: operate web pages and desktop screens via shell commands.