A macOS desktop browser designed for a human user and an AI agent to share the same application, giving each agent an isolated Space so parallel browser tasks inherit real logins without stealing the user's tabs.
Project overview
Unlike conventional frameworks that drive a separate browser, this project provides isolated parallel Spaces inside a single shared browser, preventing tab conflicts between the user and automated agents.
Project type
AI Agent · AI Coding
Use cases
Coding & Development · Automation
Deployment
Refer to project documentation
License
MIT
Best for
macOS users who need to run parallel browser automation tasks via an agent while maintaining uninterrupted access to their own browser tabs and active logins.
Key capabilities
Provides each agent with a fully isolated Space to run multiple browser tasks concurrently without colliding or stealing the user's tabs.
Acts as a connection layer between any agent CLI and the browser, exposing in-page JavaScript tools such as snapshot, fill, click, wait, navigate, and capture.
Limitations and risks
The application currently runs only on macOS; Windows and Linux support is on the roadmap.
Getting started
Setup is rated as easy and begins by downloading the macOS application.