Rob Pruzan

Pruzan’s professional profile describes earlier work on Vercel’s Next.js team, where he built an internal development environment, and identifies him as a co-author of React Scan. It lists University at Buffalo attendance from 2021 to 2025. His GitHub profile now centers Zenbu and pins the two terminal tools alongside React Scan and a small React reimplementation.[1][5]

A browser beside the agent

terminal-browser gives a developer a full web page inside a terminal tab, including a split-pane view beside a coding agent. Its command-line interface can list browser instances and let an agent act on the open page. The same workspace can display an HTML plan before implementation or preview a remote development server through SSH. This reduces the need to coordinate a terminal session and a separate browser window while checking an agent’s work.[2]

How a graphical application fits a terminal

The project renders Chromium output through Electron and sends changed image regions to terminals that support the kitty graphics protocol. Keyboard and mouse events travel back to Chromium as synthetic input. Pruzan’s README also describes a background Swift component for input the terminal does not expose. The browser interface uses a custom React renderer over a graphics engine; this is a graphical browser displayed in a terminal, rather than a text-only conversion of a page.[4]

Putting the editor in the same place

terminal-code combines code-server, which runs VS Code in a browser, with terminal-browser. Its command line can open a folder or file, compare files, jump to a line, or start with the source-control panel. It also imports settings, keybindings, snippets, and extensions from compatible editors. Together, the two projects bring editing and browser inspection into an existing terminal-based workflow without requiring the user to abandon familiar editor behavior.[3]