⌘K
Change language Switch ThemeSign In
Narrow Mode
bb-browser: A "Cheat-Level" Web Information Extraction Tool Based on CDP ========================================================================
bb-browser: A "Cheat-Level" Web Information Extraction Tool Based on CDP ========================================================================  ### yan5xu
@yan5xu
😅嗯，bb-browser，badboy browser，坏孩子浏览器来了，真的很丧良心，但真的很好用。
现在你可以用 bb-browser site 的方式直接拉到任何网站的信息，目前支持 Reddit、Twitter、GitHub、Hacker News、小红书、知乎、B站、微博、豆瓣、YouTube，50+ 个命令，我会持续更新。
当然能做到信息获取这件事不稀奇，我也是看到 @jakevin7 的 twitter-cli 的启发，才做的。但 bb-browser 的实现方式非常丧良心 — 我是通过 Chrome 插件 + CDP 直接操控你真实的浏览器。不是无头浏览器，不是偷 Cookie，不是模拟请求。你已登录了，它就直接用你的登录态。它直接在浏览器 console 里面跑 eval，以前爬虫最麻烦的登录态、还有各种鉴权都没有了😂。（这种方式真的。。。太作弊了，我都能想到哪些大厂前端发现我在这么搞，会怎么骂我，因为真的很难防）
另外我还在命令行里面埋了 guide 命令，也就是说你只要装了 bb-browser CLI 或 MCP，跟你的 Agent 说"我需要把 XX 网站 CLI 化"，它就能帮你做了！！Show More
Mar 14, 2026, 4:38 PM View on X
23 Replies
16 Retweets
130 Likes
8,515 Views  yan5xu @yan5xu
One Sentence Summary
A tool that directly controls a real browser via a Chrome extension and CDP, bypassing login and authentication to extract information from over 50 websites.
Summary
The author has released bb-browser (badboy browser), an innovative web information extraction tool. Its core technical approach involves directly controlling the user's real browser via a Chrome extension combined with CDP (Chrome DevTools Protocol), rather than using headless browsers or simulating requests. This method allows direct reuse of the user's login session and authentication information, executing 'eval' in the browser console to extract data. It currently supports over 50 sites, including Reddit, Twitter, GitHub, and Zhihu, and can be invoked by AI Agents via an MCP interface.
AI Score
88
Influence Score 88
Published At Today
Language
Chinese
Tags
bb-browser
Browser Automation
CDP
Web Scraping
AI Agent HomeArticlesPodcastsVideosTweets
bb-browser: A "Cheat-Level" Web Information Extraction To... ===============