Which AI tools do people use to let AI control your computer (computer use)?

Across the 10 sourced use cases in this guide, the most used tools are Claude Code, OpenAI Codex, ChatGPT, Arduino, Chrome.

How much does it cost to get started?

1 of the 10 use cases can start on a free plan. The most common paid starting price is $17/mo (4 use cases). Prices come from each tool's official pricing page.

How long does setup take?

3 of the 10 use cases list a setup time of an hour or less. Each card shows the setup time the builder reported.

Are these real examples?

Yes. Every use case links to its public source (an X post, a Hacker News thread or a GitHub repo) with the author and date, so you can check it yourself.

How to let AI control your computer (computer use)

Short answer: Computer use means an agent looks at your screen (or the app's accessibility tree), then clicks, types and scrolls like a person. You can use a built-in feature such as Codex computer use or Grok Bot's browser, or add an open-source computer use MCP server to Claude Code, Gemini CLI or another agent. Builders below use it to drive Mac apps in the background, fill forms, dispute fees by email and even control an iPhone through an Arduino mouse. It is slower than an API, so use it where no API exists, run it in a separate desktop or VM, and approve purchases, sends and deletions yourself.

How builders let agents see the screen, move the mouse and type, from Codex's computer use inside Claude Code to an Arduino that taps an iPhone. 10 matching use cases below, each with a public source.

Looking for something else? Search all 1241 use cases or .

Get started in 3 steps

Prefer an API or MCP integration when one exists; use computer use for apps and sites that have no other way in.
Give the agent its own browser profile, desktop or VM, logged into only the accounts the task needs.
Keep a human approval step for payments, sending messages and deleting anything, and watch the first few runs live.

Automation & businessX · @nidhisinghattri

Let Claude Code control your Mac in the background with Codex's computer use

A developer exposed Codex's computer-use ability to other agents and plugged it into Claude Code, which has no built-in way to drive a Mac from the CLI. Claude Code can now click and type in Mac apps in the background while you keep working.

$17/mo30 minutesIntermediate

Claude Code OpenAI Codex macOS

Dev & opsX · @spreen_co

Port Codex's computer-use feature to Rust for any agent

He reverse-engineered the computer-use component that ships with the open-source Codex and rebuilt it in Rust so Claude Code and other agents can use it.

$17/moDaysAdvanced

OpenAI Codex Claude Code Rust

Vibe coding & appsX · @AndreaPaparella

Give Codex control of an iPhone with a screen stream and an Arduino mouse

A developer built computer use for iPhone: a small iOS app streams the screen to a server, the server offers MCP tools to Codex, and an Arduino acting as a mouse taps the phone wherever Codex points.

Free tier1 weekendAdvanced

OpenAI Codex Arduino iPhone

Dev & opsX · @DanKornas

Add computer use to any coding agent with an open-source MCP server

open-computer-use wraps computer control (clicking, typing, reading the screen) as an MCP server for macOS, Linux and Windows. Install it with npm, then use its setup commands for Codex, Claude Code, Gemini CLI or opencode; it can run single actions or whole sequences.

$17/mo15 minutesIntermediate

OpenAI Codex Claude Code Gemini CLI

Automation & businessX · @cgarciae88

Vibe-code your own computer-control tool and use it from every assistant

The author vibe-coded a tool that controls their computer to do things current harnesses cannot, and uses it from ChatGPT, Claude, Gemini and Grok Bot. When something is missing they tell Dots to add it.

$30/mo1 dayIntermediate

ChatGPT Claude Gemini

Personal & lifeX · @GenAI_is_real

Have a computer use agent fight unfair fees with emails and complaints

After a rent payment failed through a new bank account, a renter was hit with a $50 late fee and a $25 ACH fee. He gave OpenAI's computer use agent his emails, payment confirmation, return notice, statement and lease. It asked for a manager's review, wrote to the corporate office and filed BBB and California consumer complaints.

See pricing page1 hourBeginner

ChatGPT Chrome

Dev & opsX · @sxhivs

Speed up computer use agents by skipping a screenshot after every action

A developer built arc-cua, an open-source computer use layer driven by fast decision models instead of a screenshot plus an LLM call after every click. It supports the Jev and OpenAI decisions APIs.

See pricing page1 afternoonAdvanced

Python

Vibe coding & appsHacker News · callbacked

Control your Mac with gestures from a Meta Neural Band

The author had Astra take over a 10-month-old proof-of-concept repo that reads sEMG data off the Meta Neural Band, then turned it into Kinesis: swipe between desktops, open Mission Control and turn a virtual volume knob with finger gestures.

$10 in / $50 out per 1M tokens1 dayAdvanced

GPT-6 Astra

Automation & businessX · @harkixsha

Let Claude Code control mouse and keyboard via screen vision

Built Rosply, a desktop agent that controls mouse and keyboard from screen vision and hooks into Claude Code as a global MCP server.

$17/mo1 dayAdvanced

Claude Code

Automation & businessHacker News · schuon

Add an API to legacy software with computer use

Open-sourced legacy-use, which wraps old desktop software in REST APIs by having a computer-use agent drive the UI.

See pricing pagedaysAdvanced

Common questions

How do you let AI control your computer (computer use)?: Computer use means an agent looks at your screen (or the app's accessibility tree), then clicks, types and scrolls like a person. You can use a built-in feature such as Codex computer use or Grok Bot's browser, or add an open-source computer use MCP server to Claude Code, Gemini CLI or another agent. Builders below use it to drive Mac apps in the background, fill forms, dispute fees by email and even control an iPhone through an Arduino mouse. It is slower than an API, so use it where no API exists, run it in a separate desktop or VM, and approve purchases, sends and deletions yourself.
Which AI tools do people use to let AI control your computer (computer use)?: Across the 10 sourced use cases in this guide, the most used tools are Claude Code, OpenAI Codex, ChatGPT, Arduino, Chrome.
How much does it cost to get started?: 1 of the 10 use cases can start on a free plan. The most common paid starting price is $17/mo (4 use cases). Prices come from each tool's official pricing page.
How long does setup take?: 3 of the 10 use cases list a setup time of an hour or less. Each card shows the setup time the builder reported.
Are these real examples?: Yes. Every use case links to its public source (an X post, a Hacker News thread or a GitHub repo) with the author and date, so you can check it yourself.

More how-to guides