Browser Use
Fast, persistent browser automation with session continuity across sequential agent commands. The browser stays open between commands (~50ms latency via a persistent background daemon), allowing efficient chaining of operations without restart overhead.
Browser Modes
Three primary options:
| Mode | Description | Setup |
|---|---|---|
| Headless Chromium | Default, zero configuration required | None |
| User's Chrome | Uses your Chrome with existing logins/cookies | Profile selection |
| Cloud Browser | Zero-config hosted sessions | API key required |
Core Capabilities
Navigation & State Management
Follows a straightforward workflow: launches browser, opens page, returns clickable elements with indices for interaction targeting.
Interaction Commands (20+ categories)
browser-use click 5— click element by indexbrowser-use input 3 "text"— fill a fieldbrowser-use select 7 "option"— select dropdown optionbrowser-use hover 2— hover over element- File upload, keyboard shortcuts, and more
Data Extraction
Retrieve page titles, HTML content, element text, input values, attributes, and bounding boxes through dedicated commands.
Session Persistence
Browser stays open between commands — no restart overhead, no login loss between steps in a workflow.
Authentication Workflow
For authenticated sites, use existing Chrome profiles to maintain existing session data:
# List available Chrome profiles
browser-use profile list
# Launch with a specific profile
browser-use launch --profile "Profile 1"
Essential Commands
browser-use open <url> # Navigate to a URL
browser-use screenshot # Capture current state
browser-use scroll down # Scroll the page
browser-use tab new # Open a new tab
browser-use tab switch 2 # Switch to tab by index
browser-use wait "#element" # Wait for element to appear
browser-use cookies export # Export cookies for session sharing
browser-use cookies import # Restore a saved session
Cloud Session Management
For zero-config cloud sessions and parallel execution, use the cloud browser mode with your API key. Supports local server tunneling for development workflows.
Installation
npx skills add https://github.com/browser-use/browser-use --skill browser-use
Mirrored from https://github.com/browser-use/browser-use — original author: browser-use, license: MIT. This is an unclaimed mirror. Content and ownership transfer to the author when they claim this account.