AutomateBrowser

AutomateBrowser lets an AI coding assistant — in VS Code, Cursor, Claude Code or any editor that speaks the Model Context Protocol — operate the browser you already have open. Not a fresh, empty automation browser. Yours: the tabs you have open, the accounts you are already signed in to, the extensions you already trust. There is no separate profile to set up and nothing to log in to twice. The assistant can navigate, read a page as structured text, click, type, fill a form in one pass, take screenshots, inspect console output and network activity, and report performance measurements. It works on ordinary pages without attaching a debugger, so your browsing is not interrupted by a debugging banner; the small number of advanced features that do need one ask for it explicitly. One feature, running a piece of JavaScript your assistant writes, uses Chrome's user-scripts feature. Chrome keeps that off until you turn on "Allow User Scripts" on the extension's Details page; everything else works without it. You stay in control. The extension acts only when an assistant is connected and asks it to. Its popup shows every connected assistant and which tab each one is working in, so you can see what is happening at a glance, and you can pin it to a single tab. Several editors can share one browser at once, each identified by name. Nothing is collected and nothing is sent to us. The extension talks to software running on your own computer over a local connection, and has no analytics, no tracking and no third-party services. Full detail: https://farazpawle.github.io/automatebrowser/privacy.html Open source, Apache 2.0: https://github.com/farazpawle/automatebrowser