Plugin Computeruse
БесплатноНе проверенComputer automation plugin for ElizaOS - enables AI agents to control the local machine (when supported) or a remote ComputerUse MCP server
Описание
Computer automation plugin for ElizaOS - enables AI agents to control the local machine (when supported) or a remote ComputerUse MCP server
README
What is this?
elizaOS is an open-source, local-first operating system for AI agents. Two parts:
- Eliza — the app. An AI assistant for desktop, mobile, and web: chat, voice, your messaging accounts, a personal-assistant brain, a non-custodial wallet, browser automation, and on-device models.
- elizaOS — the runtime and the OS underneath it. The same runtime can take over the whole machine — boot a Linux desktop or run on Android as the system assistant.
The agent, your data, and the models all run on your device. Eliza Cloud is optional — add it for hosted inference, sync, and deploys.
Get Eliza
| Platform | How |
|---|---|
| Web | Open app.elizacloud.ai |
| Desktop — macOS · Windows · Linux | Download from GitHub Releases |
| iOS · Android | App Store · Play · sideload |
Run from source:
git clone --filter=blob:none https://github.com/elizaos/eliza.git
cd eliza
bun install
bun run dev # API + the Eliza app UI
To run a whole device as elizaOS instead, see elizaOS — the operating system.
What Eliza does
The app ships with:
- 💬 Chat — one inbox for your agent and your accounts (iMessage, Discord, Telegram, WhatsApp, Slack, Farcaster, X).
- 🎙️ Voice — hands-free voice with on-device transcription and natural speech.
- 📇 Phone · Messages · Contacts — telephony and SMS surfaces (native on Android).
- 🧠 Personal assistant — calendar, reminders, inbox triage, tasks, daily brief, and owner-approved actions.
- 🌐 Browser & computer use — the agent drives a real browser and desktop.
- 👛 Wallet — non-custodial EVM + Solana: transfers, swaps, bridges, LP. Every spend needs your OK.
- 📄 Documents — ask questions over your files (RAG).
- 📷 Camera & vision — capture and reason over images.
- ⚙️ Automations — schedule recurring work; pick models and routing.
Private by default
Eliza can run the whole pipeline on your device via Eliza-1, the on-device model family (Gemma-4):
- Text generation and embeddings, from ~2B (phone) to ~27B (desktop).
- Voice — local speech-to-text and text-to-speech; audio never leaves the device.
- Vision & images — on-device description and generation.
Pick a model in Settings → Model Routing and it downloads and pins; from then on it works with no network.
Apps on elizaOS
elizaOS runs apps, not just an agent. An app is a plugin that adds a surface inside Eliza; the runtime installs, launches, and tracks it like real software, and it survives restarts.
- Browse & install — catalog at plugins.elizacloud.ai; curated entries in packages/registry, plus any npm package tagged
elizaos. - First-party apps — plugin-browser, plugin-documents, plugin-phone, plugin-task-coordinator.
- Earn — apps deployed through Eliza Cloud can be metered and monetized.
elizaOS — the operating system
packages/os is the real, bootable distribution. Downloads and hardware are at os.elizacloud.ai.
- Linux (packages/os/linux) — boots a full desktop with Eliza built in from a USB stick. amd64 · arm64 · riscv64.
- Android (packages/os/android) — Eliza is the system launcher and assistant, on Pixel-class devices.
The OS is bootable today; full device certification and production update channels are in progress — see the per-target READMEs for status.
Eliza Cloud (optional)
Optional managed backend for going beyond one device. Never required — local-only is first-class. It adds:
- Auth — accounts and sign-in (OAuth/SIWS).
- Hosted inference — model routing across providers, billed per use.
- Deploy — push an agent or app to a container with its own domain.
- Sync & bridge — state across devices; drive a local agent from the cloud dashboard.
- Monetization — metering and creator earnings for published apps, agents, and MCPs.
Build on elizaOS
The runtime is open source and yours to extend. Start with the CLI:
bun add -g elizaos@beta
elizaos create my-app --template project # a deployable app workspace
elizaos create my-plugin -t plugin # a runtime plugin (action/provider/service)
The runtime is model-agnostic (OpenAI, Anthropic, Gemini, Grok, Llama, local Eliza-1, …) and extended through a small set of primitives:
@elizaos/core(packages/core) — the agent loop, plugin model, and message/memory/state primitives.@elizaos/agent(packages/agent) —AgentRuntime, the plugin loader, and the default plugin map.@elizaos/app-core(packages/app-core) — the API + dashboard host that runs agents.elizaos(packages/elizaos) — the CLI:create,info,upgrade.
A plugin exports a Plugin that registers actions (what the agent does), providers (prompt context), services (long-lived singletons), and evaluators (post-response work). Import @elizaos/core directly to use the runtime with no CLI or UI — see packages/examples and the evaluation suites in packages/benchmarks. Full guides: docs.elizaos.ai.
Working in the monorepo
bun install # workspace install
bun run install:light # skip the large artifact download when disk or time is tight
bun run dev # API + Vite UI for apps/app
bun run build # turbo build across the workspace
bun run test # full test suite
bun run cloud:mock # boot the local cloud backend stack with mocks
The repo is self-contained — runtime, CLI, dashboard, native OS forks, cloud backend, and first-party plugins all live here. Every package carries its own README.md; read it before working inside. Tree map: AGENTS.md.
Contributing
Contributions welcome. Open an issue before sending a non-trivial PR. Before opening a PR, read CONTRIBUTING.md and the evidence standard in AGENTS.md; frontend-testable changes need screenshots (JPG), an MP4 video walkthrough, logs, and any relevant real-LLM trajectories attached inline in the PR itself. Agents and maintainers working from a GitHub Project board use the coordination workflow in CONTRIBUTING.md before claiming kanban work. Active MVP work is coordinated on the LifeOps Personal Assistant MVP board and in Discussions; in-flight design docs live in packages/docs/ongoing-development/.
To report a security vulnerability, follow SECURITY.md; do not open a public issue.
License
MIT — see LICENSE.
Установить Plugin Computeruse в Claude Desktop, Claude Code, Cursor
unyly install plugin-computeruseСтавит в Claude Desktop, Claude Code, Cursor и VS Code — сам разбирается с npx, uvx и сборкой из исходников.
Впервые? Поставь CLI: curl -fsSL https://unyly.org/install | sh
Или настроить вручную
Выполни в терминале:
claude mcp add plugin-computeruse -- npx -y @elizaos/plugin-computeruseFAQ
Plugin Computeruse MCP бесплатный?
Да, Plugin Computeruse MCP бесплатный — установка в пару кликов через Unyly без оплаты.
Нужен ли API-ключ для Plugin Computeruse?
Нет, Plugin Computeruse работает без API-ключей и переменных окружения.
Plugin Computeruse — hosted или self-hosted?
Self-hosted: сервер запускается локально на твоей машине командой из раздела установки.
Как установить Plugin Computeruse в Claude Desktop, Claude Code или Cursor?
Открой Plugin Computeruse на unyly.org, выбери вкладку своего клиента (Claude Desktop, Claude Code, Cursor) и нажми Install — конфиг сгенерируется автоматически, без правки JSON.
Похожие MCP
Fetch
Web content fetching and conversion for efficient LLM usage.
AWS KB Retrieval
Retrieval from AWS Knowledge Base using Bedrock Agent Runtime.
автор: modelcontextprotocolSpring AI MCP Server
Provides auto-configuration for setting up an MCP server in Spring Boot applications.
llm-analysis-assistant
A very streamlined mcp client that supports calling and monitoring stdio/sse/streamableHttp, and can also view request responses through the /logs page. It also
автор: xuzexin-hzCompare Plugin Computeruse with
Не уверен что выбрать?
Найди свой стек за 60 секунд
Автор?
Embed-бейдж для README
Похожее
Все в категории ai
