Gives your agent the ability to automate browser interactions for web testing, form filling, data extraction, and authenticated browsing using local or cloud-based sessions.
When you need to automate a multi-step workflow on a website (e.g., logging in and exporting data)
When you want to run automated E2E tests for your web application in a real browser
When you need to scrape information from a site that requires complex interactions or login sessions
Open the target URL using a Chromium or real Chrome browser mode
Run the 'state' command to identify clickable elements and their indices
Use indices to interact with the page (click, input, select, scroll)
Capture screenshots or extract data using JavaScript evaluation to verify your actions
你
Log into my GitHub account and list all my private repositories.
Agent
I've opened a real Chrome session using your default profile and successfully navigated to GitHub. I've identified the repository list elements and am now extracting the names of your private repos for you.
赋予您的代理自动化浏览器任务的能力，例如使用 CLI 导航网站、填写表单、捕获屏幕截图以及提取数据。
赋予您的代理使用 Expo DOM 组件在原生平台的 Webview 中运行 Web 代码的能力。
赋予你的智能体通过 Gemini Computer Use 模型和 Playwright 控制浏览器，从而自动化 Web 浏览器任务的能力。
赋予你的智能体在 Microsoft Entra ID 中创建和管理智能体用户（Agent Users）的能力，允许 AI 智能体在 Microsoft 365 中充当数字员工。
赋予您的智能体使用官方 rmcp SDK 生成完整、生产就绪的 Rust 模型上下文协议 (MCP) 服务器项目的能力。
