Skip to content
MoggingLabs

Docs

Features

What you can do in the app and how to do it, feature by feature. Everything on this page is shipped and current.

The multi-pane grid

One window hosts 1 to 16 live terminals in grid layouts, each pane a real PTY subprocess. Drag any divider to resize; keyboard navigation moves focus between panes; the renderer is tuned to hold ~140 fps with all 16 panes streaming.

Ctrl+T new workspace · pick a layout in the wizard · drag dividers to resize · every pane accepts any CLI

try it in the demo

Agent awareness

Every pane tracks busy, idle, and needs-you states, parsed from terminal signals rather than polling. Workspace tabs carry a count of panes waiting on you, so across ten client projects the one that needs input is findable in a second.

The needs-you badge lights the workspace tab · click it to land on the waiting pane

try it in the demo

Command blocks

Output is grouped into collapsible blocks per command, with exit-code coloring, timestamps, and search. Long agent sessions stay navigable: fold the noise, find the failure, copy the exact command that produced it.

try it in the demo

Persistent workspaces

Workspaces are SQLite-backed and restore their grid, folders, and agents after a restart. Each workspace carries its own vivid color identity so ten client projects stay tellable at a glance. Panes outlive the UI: the PTY host is a separate process, so a crash or update never kills a running agent.

try it in the demo

The launcher

The app opens on a home screen with recent workspaces (opening a folder that is already live switches to it instead of duplicating it), presets that pre-seed the wizard, and a quick terminal for one-off work.

Ctrl+T from anywhere · recents are one click

try it in the demo

Command palette

Every action in the app is reachable from the keyboard through the palette: switch workspaces, toggle the dock, launch agents, open settings.

Ctrl+K opens the palette

try it in the demo

Worktree isolation

Any pane can run its agent in its own git worktree, so parallel agents never trample each other’s files. Each worktree is a real branch checkout; the pane header shows which one.

try it in the demo

Diff review before landing

Work done in an isolated worktree goes through a diff review inside the app before it lands: see every change an agent made, approve or reject, and nothing merges without you.

try it in the demo

The task board

A Kanban board where the card is the prompt: drop a task on a lane, launch an agent with that task as context, and watch the card move as the pane works. Cards connect to worktrees and the review gate, so board-driven work lands the same safe way.

try it in the demo

Swarm coordination

For bigger jobs, panes coordinate: roles define who does what, a shared mailbox passes messages between agents, exclusive file ownership stops two agents reserving the same files, and a reviewer-gated merge keeps a human owning what ships.

try it in the demo

Profiles and usage failover

Profiles point a pane at a different CLI config home (a work account beside a personal one, for example). Credentials never move; a profile only selects which home the CLI itself reads. When a provider hits its usage limit, failover can switch the pane to the next profile.

try it in the demo

Remote SSH panes

A pane can live on another machine over SSH: same grid, same awareness, with the agent running on the remote box.

try it in the demo

The browser dock

A toggleable browser dock previews what your agents build, beside the grid rather than in another window. Each workspace remembers its last preview URL. The dock is sandboxed and runs in its own isolated session; details on the Security tab.

Ctrl+Shift+U · the titlebar globe · or the palette

try it in the demo

Agent browser control

With per-workspace consent, agents can drive the dock through a first-party MCP server: navigate, read the page as a structured snapshot, click, type, take screenshots, watch the console and failed requests. The build, preview, see-the-error, fix loop closes without a human alt-tabbing. The toolset is documented on the Automation tab.

try it in the demo

Usage meters

A titlebar gauge answers the all-day question: can I keep going? The popover shows session and weekly windows for your plan and a pace verdict, whether you will run out early, land on pace, or leave quota unused. Behind it, a full Usage tab adds a local spend scan over your own logs, usage history, provider status, and month-to-date windows, across a broad provider catalog led by Claude and Codex. How the meters read your plan without touching your credentials is on the Security tab.

Click the titlebar gauge · configure in Settings, Usage

try it in the demo

Settings

A full settings page with sectioned navigation: agents and profiles, appearance, usage, updates, and the setup checklist. Everything is stored locally.

try it in the demo