Agent Browser Automation
Automates complex web browser interactions, data extraction, and mobile simulation via a dedicated CLI for AI agents.
Agent Browser is a robust capability for Claude Code that enables seamless programmatic interaction with websites. It uses a unique snapshot-and-reference system to allow AI agents to precisely navigate pages, fill forms, click elements, and manage authentication states through CLI commands. Beyond basic automation, it supports advanced workflows like parallel sessions, session state persistence for persistent logins, and mobile environment simulation for iOS Safari, making it an indispensable tool for automated testing, web scraping, and multi-step web task automation.
주요 기능
01Interactive element referencing system (@e1, @e2) for precise DOM interaction
02Session state persistence to save and reuse authentication cookies and storage
03Comprehensive page capture tools including screenshots, PDFs, and JSON data extraction
040 GitHub stars
05Mobile simulation including iOS Safari support and touch-specific gestures
06Visual debugging features with headed browser mode, element highlighting, and video recording
사용 사례
01Complex web scraping and data collection from dynamic, JavaScript-heavy sites
02Automating repetitive administrative tasks like form submissions and report generation
03Automated end-to-end testing and visual regression checks for web applications