Get started
THE INTERFACE

What each screen is for

One line per screen: the question it answers and the call behind it.

Every screen renders from the same tools an agent calls. Where one call carries the whole page, it is named - so you can ask an agent for the same picture you are looking at.

Daily

ScreenThe question it answers
Command CenterWhat needs me, how is the portfolio moving, what changed. One get_workspace_overview call.
InboxThe single queue of things waiting on a person: approvals, open decisions, blocked gates, unresolved incidents. list_inbox.
TimelinePlan against actual across initiatives, milestones, releases and dated work.
BoardThe kanban for one product - columns, epics, tasks. get_board.

Per product

ScreenThe question it answers
OverviewThe five why fields, the current stage, health and its reasons.
CanvasThis product's slice of the timeline, centred where you clicked.
GatesEvery checkpoint, its conditions and what is unmet.
FinancialsBurn, net, runway - computed on read and explained, never stored.
InfrastructureWhat this product runs on, and the blast radius of touching it.

Across the workspace

ScreenThe question it answers
DecisionsEverything open, everything decided, and the Launch Window focus view for one call.
KnowledgeDocuments and named logs, with version history.
LibrarySkills, agents and prompts, plus the collections that bundle them.
InstructionsReusable AI rules and the sets that group them.
TeamMembers, roles and invitations.
SettingsProfile, workspace, MCP tokens, telemetry policy and infrastructure.

Two things worth knowing

Cmd+K opens the command palette. It jumps to any screen and searches decisions, instructions, library assets and infrastructure live.

Quick capture drops a thought onto the timeline in about five seconds, to be turned into a task, epic, milestone or initiative later. Its origin survives that conversion.