Browser Use Desktop
https://github.com/browser-use/desktop- Category
- AI Agents
- Rank
- No. 1334Tools index
Previous survey · No. 1340 ·
- Pricing
- Open Source
- Type
- APP
- Builder
- browser-use
- GitHub
- 676 stars
- Latest release
- v0.0.31
- Date
About
Desktop app for Browser Use — run AI browser agents on your machine with a native UI for managing tasks and tools.
What it does
Browser Use Desktop keeps Chrome as your everyday browser, then copies its cookies into a fresh Chromium instance for agent work. A global keyboard shortcut can start tasks from anywhere. Sessions run through Anthropic or Codex, while WhatsApp messages can also trigger them.
Why it's ranked here
The tool has a clear boundary: it supplies the agent layer without replacing normal Chrome. Cross-platform installers, authenticated local task submission, automated Linux artifact checks, and focused unit and onboarding tests make it more credible than a thin desktop wrapper. Provider and messaging choices remain narrow.
What's good
Cookie transfer lets agents begin with existing signed-in sessions instead of forcing repeated logins. The app ships native installation formats for three desktop platforms. Its local task endpoint uses bearer authentication, Linux builds are isolated with Docker, and crash reports scrub emails, tokens, secrets, and user paths before transmission.
Tradeoffs
Provider support is limited to Anthropic and Codex. The documented messaging channel is WhatsApp, with no other inbound services listed. macOS downloads target Apple Silicon, Windows targets x64, and the Linux AppImage targets x64. General test coverage is reported but does not yet block continuous integration.
How to use it well
It fits people who already work in Chrome and want browser agents launched quickly from the desktop, terminal, or WhatsApp. Use the fresh Chromium environment for authenticated web tasks while keeping ordinary browsing separate. It does not replace Chrome, provide a broad provider marketplace, or serve as a general messaging automation hub.
Technical notes+
app/forge.config.ts defines an Electron Forge and Vite build with separate shell, pill, onboarding, logs, and popup renderers. It packages DMG, Squirrel, DEB, and RPM outputs, reinstalls production externals, rebuilds native modules against Electron, and supports conditional signing. scripts/build-linux-appimage.mjs creates the x64 AppImage, while scripts/verify-linux-artifacts.mjs checks DEB, RPM, AppImage, and update-feed outputs. app/scripts/run-task.mjs submits prompts to a running app through a bearer-authenticated local control endpoint. app/vitest.config.ts covers unit and integration suites without global enforcement; app/vitest.onboarding.config.ts applies explicit coverage thresholds. app/config/sentry.ts conditionally enables Sentry and scrubs sensitive fields and strings. app/harnessless/daemon.js maintains a Chrome DevTools Protocol WebSocket and exposes a local IPC relay.
Observed
- License
- MIT
- Application stack
- Electron Forge, Vite, TypeScript, React, and JavaScript support scripts
- Platform support
- macOS Apple Silicon, Windows x64, and Linux x64
- Packaging
- DMG, Squirrel EXE, AppImage, DEB, and RPM distributions
- Interfaces
- Desktop interface, global keyboard shortcut, WhatsApp trigger, and local task-submission CLI
- Agent providers
- Anthropic through a Claude Code subscription or API key, and Codex through a ChatGPT subscription or API key
- Testing
- Vitest unit, integration, onboarding, identity, and React component test configurations are present
Read from README.md, app/forge.env.d.ts, app/forge.config.ts, app/eslint.config.js, app/vitest.config.ts, app/vite.main.config.ts, app/vite.preload.config.ts, app/vitest.onboarding.config.ts, scripts/build-linux-appimage.mjs, scripts/verify-linux-artifacts.mjs, app/config/sentry.ts, app/harnessless/run.js, app/harnessless/paths.js, app/scripts/run-task.mjs, app/harnessless/daemon.js.
What it can do
Execute AI browser automation tasks
Task instructions or commands → Automated browser actions and results
Manage AI browser agents
Agent configurations and settings → Active AI agents ready for task execution
Monitor browser agent activity
Running browser automation tasks → Real-time task progress and status updates
Configure browser automation tools
Tool settings and parameters → Customized automation environment
Run web scraping operations
Target websites and data extraction rules → Extracted web data
Automate web form interactions
Form data and interaction sequences → Completed form submissions and responses
Tags
Tech Stack
Media
Comments (0)
No comments yet
Editorially curated, with community endorsements as a secondary signal. Corrections welcome.