Mcp Browser Agent

by yubao2000

279 downloads
Not rated
GitHub

About

让 AI Agent(Claude/ChatGPT/Codex)像人一样操作浏览器。36个工具:截图、点击、填表、提取、PDF导出、签页管理、Cookie持久化、批量自动化。开源免费,npm 全局安装即用。

Details

Author
yubao2000
Downloads
279
Categories
AI, Automation

- 30 tools for navigation, clicking, filling forms, and scrolling
- Extract text, HTML, screenshots, and PDFs from pages
- Manage tabs: open, switch, close, and iframe contexts
- Execute custom JavaScript and retrieve console logs
- Cookie persistence for maintaining login sessions
- Optional visible browser window for debugging or manual intervention

Setting up with Highlight

This MCP is not yet compatible with Highlight’s one-click setup. However, you can still use it with Highlight by following these steps:

  1. Download and install Highlight from highlightai.com/download
  2. Navigate to the plugins tab and select "Add Custom Plugin"
  3. Configure the plugin with the settings below
    Plugin Name Mcp Browser Agent
    Command (node, npx, python, etc.)

    Please refer to the README for specific instructions on how to obtain API keys or other required environment variables.

  4. Enable "Start Automatically" if you want the plugin to start when Highlight launches

From the repository

Install the package globally via npm (npm install -g @yubao2000/mcp-browser-agent), then start the server with the command mcp-browse-agent. Configure it in supporting MCP clients (Claude Desktop, Cursor, or OpenAI Codex) by adding a new server entry pointing to the same command. A configuration file .mcp-browser.json can be used to set options like viewport, headless mode, and data directory.

Claude Desktop / Cursor

Paste into your MCP client config file to install this server.

{
    "mcpServers": {
        "mcp browser agent": {
            "browser-agent": {
                "command": "mcp-browse-agent",
                "args": []
            }
        }
    }
}

McpServers

{
    "browser-agent": {
        "command": "mcp-browse-agent",
        "args": []
    }
}

🤖 MCP Browser Agent

让 AI Agent(Claude/GPT)像人一样操作浏览器。 Let AI agents control the browser just like a human. npm version License: MIT Node ---

📋 工具清单(30 个)

🧭 导航类

| 工具 | 说明 | |------|------| | browser_screenshot | 打开网页并截图(支持全页截图) | | browser_getUrl | 获取当前页面 URL | | browser_getTitle | 获取页面标题 | | browser_getHTML | 获取 HTML 源码 | | browser_reload | 刷新页面 | | browser_back | 后退 | | browser_forward | 前进 | | browser_pdf | 导出当前页面为 PDF |

🖱️ 交互类

| 工具 | 说明 | |------|------| | browser_click | 点击元素(CSS 选择器、文本、右键) | | browser_fill | 填写输入框(模拟真实输入) | | browser_clear | 清空输入框 | | browser_select | 下拉列表选择(value/label/index) | | browser_pressKey | 键盘按键(Enter/Escape/Tab/组合键) | | browser_hover | 鼠标悬停 | | browser_submit | 提交表单 | | browser_scroll | 滚动(上/下/到顶/到底/到元素) | | browser_drag | 拖拽元素 |

📝 提取类

| 工具 | 说明 | |------|------| | browser_extract | 提取文字内容(支持选择器和属性) | | browser_evaluate | 执行自定义 JavaScript | | browser_screenshotElement | 截取指定元素的截图 | | browser_getCookies | 获取 Cookie | | browser_deleteCookies | 删除 Cookie |

📑 标签页管理

| 工具 | 说明 | |------|------| | browser_newTab | 打开新标签页 | | browser_switchTab | 切换到指定标签页 | | browser_closeTab | 关闭标签页 |

⚡ 高级功能

| 工具 | 说明 | |------|------| | browser_wait | 等待(时间或元素出现/消失) | | browser_iframe | 切换 iframe 上下文 | | browser_console | 获取控制台日志 | | browser_reset | 重置浏览器实例 | | browser_close | 关闭当前页面 | ---

🚀 快速开始

前提

- Node.js >= 18 - (推荐)Claude Desktop / Cursor / 任何支持 MCP 的客户端

快速使用

```bash
No reviews yet — be the first

Sign in to leave a review

Use Google, GitHub, or an email account so ratings stay tied to real people.

Email sign in

No reviews posted yet.