Desktop App Tour

Desktop App Tour

The desktop app is not where you do the work by hand. It is where you give the agent its hands and eyes, watch what it does, and feed it evidence. This page is a one-time orientation: open it once, learn the layout, then mostly drive through your AI agent.

Four Areas

The window has four areas. Everything you do lives in one of them.

1. Sidebar (left)

Your projects and services live here.

• Add project picks a directory that contains .superdev/config.yaml.
• Services are grouped by environment (env group, e.g. dev, prod). Each group has Start all / Stop all and a per-service start / restart / stop.
• Search services filters the list.
• Drag a service into the workspace to open its logs in a panel — this is the main way you open logs.
• At the bottom: Node Center, Settings, and the Getting Started entry.

2. Workspace (center) — tabs

The center is a multi-tab workspace. A tab is one of four kinds:

Mark	Tab	What it shows
▣	Project	Live logs for the services you opened, in split panels
◉	Overview	Project runtime status, pipelines, config, ingress
◎	Node Center	Health of all remote nodes
⌕	Search	Cross-service log search and trace context

Tabs can be opened, activated, and closed independently, so you can keep logs, a search, and the overview side by side.

3. Bottom bar

The bottom bar is the runtime status line and the place for whole-session actions:

• Panel services and open deployments currently shown.
• Runtime status, restart / stop on the selected deployment.
• Evidence (pins): pin log lines, copy, and export the bundle.
• Agent and MCP connection state, and the Approvals entry.
• Log display scope: which services and how many nodes are in view.

4. Menu-bar popover

SuperDev also lives in the system menu bar. The popover is a quick remote: search services, batch start / stop, and jump to logs without opening the main window.

You Drive Through The Agent

The fastest path is not clicking every panel — it is telling your AI agent what you want. The desktop app ships a Getting Started flow (sidebar entry) with a main line of steps, each carrying a ready-to-copy sentence to say to the agent:

1. Connect your AI (install MCP).
2. Run the sample approval loop.
3. Onboard your own project and start services locally.
4. Add a remote host and install the agent in Node Center.
5. Create a prod environment and manage remote services.
6. Run a pipeline deployment (optional).

Use the panels in this section to watch and verify what the agent does — not to replace it.

Next

• Services and the log panel — open logs, split panels, filter.
• Log investigation techniques — filter rules, evidence pins and segments, cross-service search.