# desktop

> desktop is an open-source agent for automation. Desktop app that runs a team of browser agents using your existing logins. It is available on macOS, Windows, Linux and whatsapp.

- **Kind:** Agent
- **Pricing:** Open source
- **Platforms:** macOS, Windows, Linux, whatsapp
- **Website:** http://browser-use.com
- **Repository:** https://github.com/browser-use/desktop
- **GitHub stars:** 719
- **License:** MIT
- **Categories:** [automation](https://ai-powered.com/categories/automation)
- **Listing updated:** 2026-11-18

## About

The README describes this as the agent half of an AI browser: it leaves your regular Chrome alone and instead copies your cookies into a fresh Chromium instance so agents can operate already signed-in sessions. Tasks can be launched from anywhere on the desktop through a keyboard shortcut.

The app is built on Browser Harness, the same browser-control project the vendor promotes for Claude Code and Codex. It connects to either Anthropic (Claude Code subscription or API key) or Codex (ChatGPT subscription or API key) as the model provider, so model costs follow whichever account you already use.

It suits people who already do their work in Chrome and want agents to act inside those accounts on their own machine. Installers are published for macOS on Apple Silicon, Windows x64, and Linux as AppImage, .deb and .rpm, and the project is MIT licensed.

## Features

- **Cookie import into Chromium:** Copies existing browser cookies into a separate Chromium so agents start out logged in to your accounts.
- **Keyboard shortcut launch:** Spawns a new agent task from anywhere on the desktop with a global keyboard shortcut.
- **Browser Harness foundation:** Runs on the Browser Harness project for controlling Chromium from agent sessions.
- **Choice of model providers:** Works with Claude Code or API keys from Anthropic, and with ChatGPT or API keys from Codex.
- **WhatsApp message channel:** Lets a WhatsApp message to yourself trigger agent sessions and receive replies.
- **Auto-updating Linux build:** The Linux AppImage supports in-app automatic updates.

## Good for

- Running several browser agents in parallel on a personal computer
- Having agents work inside logged-in web apps without re-entering credentials
- Starting a browser task quickly from any window with a shortcut
- Triggering agent work by sending a WhatsApp message

## FAQ

### What license is the desktop app under?

The repository is licensed under MIT.

### Which operating systems are supported?

The README lists macOS on Apple Silicon, Windows x64, and Linux, with AppImage, .deb and .rpm packages.

### Which model providers can it use?

Anthropic, through a Claude Code subscription or API key, and Codex, through a ChatGPT subscription or API key.

### How do I install it from source?

The README says to install Task (brew install go-task) and run task up to install dependencies and start the app.

### Is desktop free?

Yes. desktop is open source, so you can use and self-host it for free.

### What are the best alternatives to desktop?

The closest alternatives to desktop are Vercel and Aside.

## Where this comes from

- Description: written from the project's README and the project's website.

## Alternatives

- [Vercel](https://ai-powered.com/agents/vercel-2): Vercel is an open-source agent for automation.
- [Aside](https://ai-powered.com/agents/aside): Aside is a freemium hosted agent for automation.

---
Source: https://ai-powered.com/agents/desktop
