computer-agent

suitedaces活跃维护

用 AI 通过终端、浏览器、鼠标和键盘操控整台电脑的桌面应用，让 Claude 等模型直接替你执行 GUI 操作。

Desktop app to control your computer with AI using your terminal, browser, mouse & keyboard

暂无版本712star89fork同步于 2026-11-07 07:31 UTC

电脑操控桌面应用 GUI自动化浏览器Agent Rust

对比 GitHub 官网 ai ai-tools anthropic claude-3-5-sonnet claude-4-5-opus

怎么上手

由 README 提炼，执行前请确认

原文

npm install
echo "ANTHROPIC_API_KEY=your-key-here" > .env
npm run tauri dev

硬件要求按 README 说明

README 未说明显存或 GPU 要求，仅列出需 Rust & Cargo、Node.js & npm 和 Anthropic API key；macOS 需授予辅助功能权限。

下载：源码 zip·GitHub Releases（安装包）本站不镜像安装包

项目说明

由 README 自动翻译提炼

没人工点评？写下你的经验

本地桌面 AI 代理，用自然语言指令替你操作鼠标键盘、截图和终端

用 AI 通过终端、浏览器、鼠标和键盘操控整台电脑的桌面应用，让 Claude 等模型直接替你执行 GUI 操作。

核心能力

截图感知屏幕后直接控制光标与键盘
提供接管屏幕的 Computer Use 模式
后台模式经 Chrome DevTools Protocol 做网页自动化
支持 Haiku、Sonnet、Opus 模型在界面切换
内置快捷键语音触发与一键停止代理

容易踩的坑

必须自备 Anthropic API key 并写入 .env，否则无法运行；macOS 首次运行需在系统设置中手动开启辅助功能权限；Computer Use 模式会真实接管鼠标键盘，运行期间不要操作电脑。

内容由模型根据 README 生成，可能不完整或不准确，以官方文档为准。看原文 README

用户点评

登录用户的真实使用记录，可按版本和硬件筛选

共 0 条

还没有人点评。第一个说出真实体验的人最有价值。

README 原文

未翻译

在 GitHub 打开

Taskhomie (Computer Use AI Agent)

Local AI agent that controls your computer. Give it natural language instructions and watch it take screenshots, move your mouse, click, type, and run terminal commands.

Built with Tauri, React, and Rust.

Demo

https://github.com/user-attachments/assets/8edd92a7-7d3e-472a-9e48-3b561f0257d6

Here, I used it to autonomously read and reply to tweets, lol. This is purely for demonstration/research, you should not attempt to do the same, lol.

Modes

Computer Use Mode - Takes over your screen. Sees what you see via screenshots and controls your cursor and keyboard directly. Use when the task spans multiple apps or needs full desktop access. You step away while it works.

Background Mode - Runs async while you do other things. Uses Chrome DevTools Protocol for web automation and terminal for everything else. Doesn't touch your mouse or keyboard. Faster and more reliable for web + terminal tasks.

Setup

Requirements:

Rust & Cargo
Node.js & npm
Anthropic API key

# install deps
npm install

# add your api key
echo "ANTHROPIC_API_KEY=your-key-here" > .env

# run dev
npm run tauri dev

# or build for production
npm run tauri build

On macOS, you'll need to grant accessibility permissions when prompted (System Settings → Privacy & Security → Accessibility).

Shortcuts

⌃⇧C - push-to-talk → computer use mode
⌃⇧B - push-to-talk → background mode
⌘⇧H - help mode (screenshot + quick prompt)
⌘⇧S - stop agent

Stack

Frontend: React, TypeScript, Tailwind, Zustand, Framer Motion
Backend: Rust, Tauri 2, Tokio
Models: Haiku, Sonnet, Opus (selectable in UI)

Contributing

PRs welcome. Hit me up on Twitter @ishanxnagpal.

License

Apache License 2.0