Capabilities
A terminal built to be watched.
Loom's terminal renders through WebGL with xterm.js on native PTYs. Quick enough to keep up with six agents streaming at once, comfortable enough to be your daily driver.
Daily driver
The fundamentals, done properly.
Tabs and splits
Open as many native PTYs as the work needs, split them side by side, and shape the layout around the task instead of the other way round. The tabs and splits guide covers the moves.
Search and true color
Search scrollback to find the one line that matters. True color output means your prompt, your tools, and every Loom theme render exactly as designed.
WSL per tab on Windows
Each tab chooses its own world: run one tab in WSL and the next on the host shell. No global switch, no separate app. Details in the WSL guide.
Part of the workspace
Streams in the background, previews up front.
Terminals keep streaming when you look away, so a fleet session never goes quiet just because you switched tabs to read a diff. And when a dev server starts anywhere in the workspace, Loom auto-detects it and offers a live web preview without leaving the app.
# run your dev server in any tab
$ npm run dev
# loom detects the server and offers a web previewQuestions
Straight answers.
Is this the same terminal the fleet runs in?
Yes. The six Claude Code sessions of the fleet live in this same terminal, each on its own native PTY. You can watch any of them or click in and type, and your own tabs sit right alongside.
Why WebGL rendering?
Six agent sessions can produce a lot of text at once. Rendering xterm.js through WebGL keeps scrolling and streaming output smooth at that scale, including in the terminals you are not currently looking at.
Are these real shells?
Every tab is a native PTY running your real shell, so your tools, your dotfiles, and the full Claude Code TUI behave exactly as they do anywhere else. On Windows you can pick WSL per tab.
Hand it the work.
Walk away.
macOS, Linux, and Windows. Around 13 MB. Free and open source.