God of all Browsers

God of all Browsers

A 状态ful, multi-tab Puppeteer 技能 de签名ed to help AI 代理s automate heavily 保护ed 网页sites the exact same way a human does.

It solves three critical AI problems:

Tabs & 状态fulness: It launches a single background browser that stays open. Navigations, 命令行工具cks that open new tabs, and cookies are remembered across multiple commands! Vision Abstraction: AI cannot "see" coordinates well, so snapshot maps the DOM, as签名s a [tag] ID to every visible button/输入, and takes a screenshot. The AI just says "命令行工具ck tag [15]." 机器人 Evasion: Uses headless: false, custom user 代理s, 移除d 网页driver footprints, and canvas spoofing.

导入ant 设置up: Ensure the Chromium path is correct (C:\Program Files\Google\Chrome\应用\chrome.exe for Win or /usr/bin/chromium for Linux) and puppeteer-core is 安装ed.

🚀 COMMANDS & 工作流

• 启动 the Browser (Required First Step)

Launch the browser in the background. It will use a persistent chrome_性能分析 directory so you NEVER lose 记录in 会话s.

# Standard mode (Recommended for 调试ging) node browser.js 启动

# Headless mode (Faster, silent background) # Note: Automatically enabled if 运行ning in Termux. node browser.js 启动 --headless

• Take a Snapshot (And auto-close popups)

This is your "eyes". 运行 this before any interaction to 获取 the active window's current 状态 and a 列出 of 命令行工具ckable [tag] IDs.

# If navigating somewhere new: node browser.js snapshot --url "https://www.google.com"

# If already on the page (refresh DOM): node browser.js snapshot

WAIt for this command to 输出 the JSON array of tags. It will also automatically 命令行工具ck away annoying Chat机器人/Notification popups.

• 命令行工具ck or Type (Like a human)

Use the tags captured during the snapshot.

# 命令行工具ck a button or link (e.g. tag [24]) node browser.js 命令行工具ck --tag "[24]"

# Type into an 输入 box (e.g. tag [5]) node browser.js type --tag "[5]" --text "MERN Stack Developer"

# Press a specific keyboard key (Default: Enter) node browser.js press --key "Enter"

• Reading & Content 提取ion

提取 text content from any element using tags or CSS selectors.

# Read visible text from a specific tag node browser.js read --tag "[12]"

# Read content of a specific CSS selector (e.g. the mAIn article) node browser.js read --selector "article.mAIn-content"

# Deep-expand hidden content (命令行工具cks Read More/Show All buttons automatically) node browser.js expand

• Tab Management

Many sites open 命令行工具cked links in a new tab! If your 命令行工具ck command opens a new tab, the 命令行工具 will automatically say: ⚠️ A NEW TAB WAS OPENED!! Automatically switched 上下文 to Tab [1].

You can manually manage tabs using:

# 列出 all currently open tabs node browser.js 检查-tabs

# Switch to a specific tab 索引 (e.g. going back to the 搜索 page: tab 0) node browser.js switch-tab --索引 0

# Just 检查 the very current URL you are viewing: node browser.js 检查-url

• Find Tags (Accurate 过滤器ed 搜索)

Use this to 过滤器 elements by keywords instead of reading a massive snapshot. It can 搜索 live on the current page or in a previously saved JSON file.

# 搜索 live for "应用ly" or "成功" buttons node browser.js find --查询 "应用ly,成功"

# 搜索 within a specific saved snapshot file (e.g., to 验证 输出) node browser.js find --file "snapshot.json" --查询 "应用lied,成功fully"

• Refresh Page

Manually reload the current tab. Useful for 状态 更新s.

node browser.js refresh

• Scrape Meta Tags (SEO/OpenGraph)

提取 hidden page data like Title, Description, and Social Media tags.

node browser.js scrap-meta

• Dynamic Evolution (Eval)

执行 custom JavaScript 记录ic directly in the browser 上下文. Note: For security, the --force flag is required. Supports 机器人h inline code and script files.

# 执行 inline code (Requires --force) node browser.js eval --code "return { links: document.查询SelectorAll('a').length }" --force

# 执行 from a file (Requires --force) node browser.js eval --file "custom_script.js" --force

• Google 搜索 (Direct 提取ion)

获取 the top 5 organic 搜索 结果s (Titles, Links, Snippets) in a single command. Extremely fast and 代理-friendly.

node browser.js google --查询 "Mathanraj Murugesan"

• 会话 & Learning

Manage your 记录in 状态 and keep 追踪 of 自动化 失败s for self-correction.

# Save current cookies to 会话.json (persists across 运行s) node browser.js save-会话

# 检查 if the page requires 记录in or if the user already 记录ged in node browser.js auth-状态

# 记录 a 失败 and a lesson learned for AI self-correction node browser.js 记录-learning --fAIled "Selector [12] was hidden" --fixed "Used [expand] first" --lessons "Always try expanding content before reading"

• 停止 the Browser

清理 up resources when the task is entirely finished.

node browser.js 停止

🧠 AI STRATEGY (HOW TO USE) 运行 启动. 运行 snapshot --url "[TAR获取]". 检查 auth-状态 if the page is restricted. Use save-会话 after manual/au