LibreAI / KI-Agenten / open-codex-computer-use
// agents · iFurySt/open-codex-computer-use

open-codex-computer-use

open-codex-computer-use ist ein Open-Source-Projekt aus der Kategorie KI-Agenten von iFurySt (iFurySt/open-codex-computer-use auf GitHub). Es hat 2.358 Sterne auf GitHub und 208 Forks, wird unter der Lizenz MIT veröffentlicht und wurde zuletzt am 7. Oktober 2026 auf GitHub aktualisiert.

Open-source Computer Use as an MCP service: let any agent drive macOS, Linux, and Windows desktops.

★ 2,4 Tsd. Sterne⑂ 208 ForksMITaktualisiert Okt 2026
▦ Visual — das README dieses Projekts enthält keinen einbettbaren Screenshot,
darum hier ein Platzhalter. Die Anleitung unten ist trotzdem vollständig.
// projektdaten

Projektdaten

GitHub-Sterne
2.358
Forks
208
Lizenz
MIT
Zuletzt auf GitHub aktualisiert
7. Oktober 2026
Kategorie
KI-Agenten
Eintragstyp
Vollständige Installationsanleitung
// so funktioniert es

So funktioniert es

  1. Clarify your goal

    Decide which agent client (Codex, Claude Code, Gemini CLI, opencode, or your own MCP client) will drive the desktop.

  2. Get the project

    Install the open-computer-use package globally from npm; it also provides the shorter ocu alias.

  3. Install and configure

    On macOS, launch it once and grant Accessibility and Screen Recording permissions; then install it into your agent using the matching install command or a manual MCP config entry.

  4. Run your first task

    Restart or reconnect your agent and ask it to perform a desktop task, or test a single tool call from the CLI first.

// funktionen

Warum es so beliebt ist

MCP-native Computer Use service

Exposes computer-use tools over MCP so any compatible agent or client can list apps, inspect app state, and act on the desktop.

One-command agent installers

Built-in commands write the right configuration for Codex, Claude Code, Gemini CLI, and opencode, plus a generic installer for other harnesses.

Cross-platform desktop control

Runs on macOS, Linux, and Windows, building on OS accessibility features rather than purely visual clicking.

Direct tool calls and scripting

Call a single Computer Use tool from the CLI, evaluate JavaScript against the runtime, or keep a REPL session with reusable element state.

Sequenced action batches

Run a sequence of calls in one process so element indexes stay valid between steps, with configurable pauses.

Health and capability checks

A doctor command checks permissions and setup, and a capabilities command reports runtime availability as JSON.

Advertisement
// selbst ausführen

Führe es selbst aus

  1. Clone the project (optional, for source use)

    Most users install from npm, but the source checkout is here if you want it.

    git clone https://github.com/iFurySt/open-codex-computer-use.git
    cd open-codex-computer-use
  2. Install the CLI from npm

    This provides both open-computer-use and the ocu alias.

    npm i -g open-computer-use
  3. Install into your agent and test

    Write the MCP config for Codex automatically, then try listing apps.

    open-computer-use install-codex-mcp
    open-computer-use call list_apps
// anwendungsfälle

Was du damit bauen kannst

Advertisement
// häufige fragen

Häufige Fragen

Who maintains it?

It is maintained by iFurySt as an open-source project inspired by OpenAI's Codex Computer Use.

Is it free to use?

Yes, it is MIT-licensed and free. You bring your own agent and model access through whichever MCP client you connect.

Does it need a paid API key?

The tool itself needs no API key; the agent driving it (for example Codex or Claude Code) uses your existing account or key for that service.

How does it compare to a well-known alternative?

OpenAI's Codex Computer Use is a closed, product-integrated feature; this project offers similar computer-use ability as an open MCP service you can attach to many different agents.

How do I get started?

Run npm i -g open-computer-use, grant permissions on macOS, run the install command for your agent, and reconnect the agent.

// weiter entdecken

Ähnliche Tools

KI-Agenten

agent-desktop

Reliable computer use for any agent: drive real apps through accessibility trees, not pixel guessing.

★ 1,8 Tsd. SterneApache-2.0
Anleitung ansehen →
KI-Agenten

ESP32 iPhone Agent

ESP32 iPhone Agent — let an AI model use a real iPhone with a $7 ESP32: AirPlay eyes, USB HID hands, no computer, no jailbreak

★ 2,4 Tsd. SterneMIT
Anleitung ansehen →
KI-Agenten

agent-world-model

Snowflake's Agent World Model: synthesise 1,000 executable MCP tool-use environments for agentic RL.

★ 465 SterneNOASSERTION
Anleitung ansehen →
KI-Agenten

klavis

An MCP integration platform that lets AI agents use tools reliably at any scale.

★ 5,8 Tsd. SterneApache-2.0
Anleitung ansehen →
KI-Agenten

osaurus

Osaurus — own your AI: a native macOS agent harness with any model, memory and fully offline execution.

★ 8,0 Tsd. SterneMIT
Anleitung ansehen →
KI-Agenten

UFO

Microsoft's GUI agent that grew into UFO³ — orchestrating agents across Windows, Linux and mobile devices.

★ 10,0 Tsd. SterneMIT
Anleitung ansehen →
// alternativen

Alternativen im Bereich KI-Agenten

Die anderen Projekte der Kategorie KI-Agenten mit den meisten Sternen im LibreAI-Katalog — die realistische Auswahl, falls open-codex-computer-use nicht das Richtige ist.

// mehr in agents

Mehr in KI-Agenten

Verwandte Projekte der Kategorie KI-Agenten im Katalog, die open-codex-computer-use bei den Sternen am nächsten stehen — zum Weiterstöbern auf beiden Seiten.

Hol dir open-codex-computer-use

Kostenlos und offen — hol dir den Quellcode, gib dem Repository einen Stern und leg los.