BrowserAct Web Automation
Automates browser navigation, JavaScript-rendered data extraction, and form workflows with built-in human-in-the-loop safety controls.
BrowserAct is a CLI-driven browser automation skill designed for AI agents operating within environments like Claude Code, Cursor, and Windsurf. It equips agents to handle complex web workflows beyond basic HTTP fetchers, including JavaScript-heavy SPAs, authenticated multi-browser sessions, interactive form submissions, DOM content extraction, and HAR network recording. With built-in Confirmation Gate protocols, sensitive operations like logins and file uploads require explicit human approval, while local-first data isolation ensures sensitive session tokens and browser profiles never leave your machine.
Key Features
01Confirmation Gate protocols ensuring user approval for browser creation and sensitive operations
02Isolated multi-browser session management for parallel workflows across multiple accounts
03Lightweight JS-rendered extraction as an advanced replacement for curl and basic fetch tools
04Full DOM interaction capabilities including form automation, file uploads, clicks, and screenshots
05Network inspection supporting XHR, fetch, and complete HAR response capturing
060 GitHub stars
Use Cases
01Extracting dynamic, client-rendered web content from single-page applications (SPAs)
02Automating multi-step web workflows like authenticated forms, testing suites, or admin dashboards
03Visually inspecting layouts and capturing screenshots to verify frontend implementations