Everything you need around the game.

The cost of five windows

Unity with agents usually means an editor, a terminal, an IDE, a browser and the console. Here they sit around the running game in one window.

Read and fix code where the game runs

• A .gitignore-aware file tree that keeps Library/ out of the way.
• A code editor with a git blame gutter, in as many panes as you want.
• Search and replace on ripgrep, streaming results as they're found, plus fuzzy quick-open.

It's also Unity's External Script Editor, so double-clicking a script in Unity opens it here.

The Unity console, in the app

Counts, per-player filtering and search over every Debug.Log line, with each clone filtered separately.

Panels for the rest of the job

The new-session menu also opens these panels, which split, tab and move like any other:

• Unity Player and Unity Console.
• Source Control, ported from VS Code.
• The Tuner: Unity's Inspector and a scene view on the objects you pick.
• Scratchpads: side projects at your Unity version, each streamed.

Open several windows to spread projects across monitors, and move a workspace between them.

Knowing what your machine is doing

Live CPU and memory for every terminal, Unity instance and internal process show a runaway agent or a wedged editor before the fans do.

The status bar shows Claude, Codex and Kimi usage limits with reset times.

Alerts, for when you walk away

A finished turn flashes the tab green with a sound, a permission stop amber. Nine sounds, including the Half-Life HEV beeps.

One screen to start

Pick a folder, terminals and agents (down to model and effort for Claude Code), the Unity mode and an optional git worktree, whose Library is a copy-on-write clone on macOS. Then press Launch.