terminal-browser is a real Chromium-based browser that renders inside a terminal emulator, written in TypeScript under the MIT licence, for developers who live in the terminal and for coding agents that need to see and operate the web.

What it is

terminal-browser is a command-line program that draws the pixels of a Chromium browser engine directly into a terminal window. It relies on the kitty graphics protocol, which terminals such as ghostty, kitty, cmux and vscode support, to display pixels produced by Chromium. Pixels are read from the GPU through Electron's offscreen rendering API, which the project states allows rendering without dropped frames. The browser chrome around the page is built with a graphics engine written in Rust, and the interface itself is defined in React through a custom renderer, so the outer UI is authored in TypeScript. Both the browser UI and the page content are drawn to the same shared canvas in the Rust engine, which is what allows interface elements to be layered over the page.

The concrete problem it solves is that a terminal session and a graphical browser are normally separate windows on separate machines, which breaks the loop between an agent working in a terminal and the websites it needs. terminal-browser replaces the separate GUI browser window with a browser running in the same terminal tab as the agent. It listens for mouse clicks, mouse position and keyboard events from the terminal and sends synthetic events to Chromium, reading events that the terminal cannot supply directly from the operating system through a background Swift app. That input path is what enables smooth scrolling and trackpad handling, including sites with infinite canvases.

Key capabilities

• terminal-browser launches the browser in the current terminal, terminal-browser open opens it at a URL, and terminal-browser --split right opens it in a split pane to the right.
• terminal-browser open --ssh performs all network requests through a remote server, and terminal-browser ls lists open browsers.
• terminal-browser action is an agent-browser compatible CLI for interacting with open terminal-browsers, giving an agent programmatic use of the web.
• terminal-browser upgrade upgrades to the latest version, and installation is available through the curl -fsSL https://terminal-browser.sh/install | bash script or brew install terminal-browser.
• DevTools open with f12 or ctrl+shift+i, with a dedicated DevTools console, and a command palette is bound to cmd+p on macOS and ctrl+k or alt+k on Linux.
• Page recording starts and stops with ctrl+r on macOS or ctrl+shift+r on Linux, review completes with ctrl+enter, and ctrl+g starts element selection for sending to an agent.
• The underlying rendering logic is abstracted into a JavaScript library at https://github.com/zenbu-labs/pixel for building graphical applications in the terminal.

Who uses it and how

• Developers who keep a coding agent and a website scoped to the same terminal tab, which the README lists as a primary use case.
• Agent workflows built on the listed claude-code, claude-code-plugin, claude-skills, codex and herdr-plugin topics, where the agent gains the capability to use the web by interacting with open terminal-browsers.
• Agent-generated HTML plans, which open inside terminal-browser in a split pane next to the agent that produced them.
• Remote development over SSH, where terminal-browser --ssh lets a developer preview websites running on remote machines without running the browser directly on the remote host.

Getting started

Install with the curl script curl -fsSL https://terminal-browser.sh/install | bash on macOS or Linux, or with brew install terminal-browser. A Claude Code plugin install path is documented at /claude-code-plugin/README.md, and the project homepage is https://terminal-browser.com.

How it compares

No comparable or paid products are named anywhere in the facts supplied for this entry, so terminal-browser stands alone in this registry. The only reference points the project itself gives are the terminals it depends on, namely ghostty, kitty, cmux and vscode as kitty graphics protocol implementations.

When to use it — and when not to

terminal-browser requires a terminal that implements the kitty graphics protocol; on terminals without it, the rendering path does not work at all, so this is not a drop-in browser for every shell. There is no server, database or mail service to operate, since installation is a script or a Homebrew formula and terminal-browser upgrade handles updates, but the project is young in release terms and much of its input behaviour depends on a background Swift app reading operating system events. Anyone needing a fully documented, platform-neutral browser should look elsewhere first, since the README leaves parts of the SSH path and the truncated notes on running the browser directly on a remote machine incomplete.

curl -fsSL https://terminal-browser.sh/install | bash

brew install terminal-browser

terminal-browser # launches the browser
terminal-browser open <url> # opens the browser at a url
terminal-browser --split right # opens the browser in a split pane to the right
terminal-browser open --ssh <user@host> <url> # performs all network requests through a remote server
terminal-browser ls # lists open browsers
terminal-browser action # an agent-browser compatible cli for interacting with open terminal-browsers
terminal-browser upgrade # upgrade to the latest version