This tool acts as an MCP server, providing AI agents with comprehensive capabilities to interact with web pages. It allows agents to navigate to URLs, click elements, type text, scroll pages, take screenshots, inspect the Document Object Model (DOM), and read console logs. It supports both headless browser automation via Puppeteer and control over a real Chrome browser via an extension, offering features like session persistence, custom viewports, and anti-bot bypass. This robust control enables AI agents to perform complex web-based tasks, from filling forms and logging in to testing layouts and debugging JavaScript.
Key Features
01Read DOM for structural analysis and console logs for debugging
02Full browser control: navigate, click, type, scroll
03Support for custom viewports and session persistence
04Screenshot pages for visual feedback
050 GitHub stars
06Headless Puppeteer mode and real Chrome browser extension mode
Use Cases
01Debugging web applications by accessing console logs and inspecting the DOM directly
02Testing website responsiveness and appearance across various mobile and desktop viewports
03Automating web tasks like form submission, login flows, and interactive workflows
