Features
One guide per view — every surface of Command Center, what you do there, and how the views fit together.
Command Center is a set of views — one focused surface per capability — inside a single desktop shell. This section documents them one by one: what each view is for, how it is laid out, and the concrete flows you run in it. Concept pages explain the ideas behind the surfaces; these pages explain the surfaces themselves.
How these guides work
Each guide follows the same pattern:
- What the view is for — the job it does in your daily operation.
- The layout — panels, tabs, and controls, in the order you meet them.
- The flows — step-by-step: what you click, what happens, what to check.
- Cross-links — the concept page that explains the why, and the reference page that owns the exact facts (configuration keys, check catalogs).
Where a view is young or deliberately minimal, the guide says so. Nothing here describes capabilities the product does not ship.
The shell around every view
A few shell behaviors apply everywhere and are worth knowing before you read any single guide:
- The sidebar is the map. Views are grouped into toned sections — Mission Control, Workflows, Data, Agents — with Terminals and Voice pinned above them and Settings at the bottom. The footer also carries your account state and a quick theme switch.
- Views your plan does not include stay on the map. They carry a lock in the sidebar and open onto a panel naming what they need, so the map is the same shape whatever your plan. See Tiers: Pro & Ultra.
- Almost every view pops out. The ⧉ button in the title bar (or hovering a sidebar item) opens the active view in its own window. Open windows and their positions are restored when you restart the app. Settings is the one view that stays in the main window — it is the single place configuration is written.
- The title bar is the header. It carries the app branding, the sidebar collapse toggle, dictation and Voice controls, and — for daemon-backed views — a context-aware status strip that swaps as you navigate.
- Cmd/Ctrl-K jumps anywhere. The command palette searches your documents, graph entities, workspaces, context artifacts, and memory in one box and navigates straight to the hit. See keyboard shortcuts.
- A broken view never takes the app down. If a view hits a render error it degrades to an in-place error card; the sidebar and title bar stay live and navigating away resets it.
The view map
| View | What you do there | Guide |
|---|---|---|
| Terminals | Run and watch agent sessions in a split-pane terminal workspace | Terminals |
| Dashboard | Read the state of the whole system at a glance; control the daemons | Dashboard |
| Voice | Talk to the platform assistant | Voice |
| Missions Board | Track missions and operations against the strategic files | Missions Board |
| Calendar | Put work on a moment in time and let an agent do it when that moment arrives | Calendar |
| Read and write mail, triage the queue, approve the agent's drafts | ||
| Identity | Edit the company identity files every agent writes from | Company Identity |
| Routines | Schedule, run, and audit recurring agent work | Routines |
| Observer | Manage news harvesting — sources, batches, runs | Observer — harvesting |
| Intelligence | Generate and review briefings and publications | Intelligence — publishing |
| Products | Drive the knowledge-product pipeline from research to export | Products |
| Video | Brief, review, and render video from an HTML-first engine | Video |
| Cortex | Browse and curate the knowledge graph | Cortex |
| Context Artifacts | Compose the context packages agent sessions launch with | Context Artifacts |
| Knowledge Base | Search and manage ingested content | Knowledge Base |
| Memory | Browse and curate what the agents remember | Memory |
| Documents | Read, edit, and organize the documents tree | Documents |
| Agents | Inspect and edit the agent roster | Agents |
| Protocols | Read, edit, and author the doctrine files that govern how work is done | Working in the Protocols view |
| Skills | Inspect the packaged skills agents invoke | Skills |
| Settings | Configure everything — the single write surface | Settings |
The Protocols row is the one guide that lives outside this section: the view is a reader and editor over the doctrine corpus, so it is documented on the doctrine page beside the files it opens.
Three more system surfaces render full-bleed inside Settings, as tabs rather than sidebar entries. Each has its own guide because each is a full view:
| Surface | Where | Guide |
|---|---|---|
| Doctor | Settings ▸ Doctor | Doctor |
| MCP Servers | Settings ▸ MCP Servers | MCP Servers |
| Data Layer | Settings ▸ Data Layer | Data Layer |
The Activity Log is a full view too, but it is reached from Dashboard ▸ Live ▸ Log rather than from the sidebar; only its retention setting lives in Settings.
Where to start
If you are new, read three guides in this order: Dashboard (the daily home surface), Documents (where the work accumulates), and Settings (how the system is wired). The operator model concept page ties them together.
Then read Autonomy & controls before you leave an agent working unattended. It is the one page that states how far a dispatched agent may reach, which switch stops each outward action, and what is recorded — the feature guides point at it rather than restating it.