---
title: Dashboard
description: The workspace hub. Recent runs, activity over time, fleet lanes per project, and where usage stands against limits.
sidebar:
  order: 4
---

The dashboard at `/dashboard` is the workspace's home. It answers "what has been
happening, and where do we stand" at a glance.

## What it shows

<CardGroup cols={2}>
  <Card title="Recent runs" icon="history">
    A table of recent runs across the organization, each a link into its session.
  </Card>
  <Card title="Activity" icon="calendar-days">
    A contributions heatmap and a runs trend, so cadence is visible over time.
  </Card>
  <Card title="Fleet lanes" icon="columns-3">
    Per-project lanes showing runs grouped by the repository or project they act
    on.
  </Card>
  <Card title="Usage and limits" icon="gauge">
    A limits card summarizing model burn and the sandbox footprint against
    available headroom.
  </Card>
</CardGroup>

The usage figures come from real run accounting rather than estimates, so the
limits card reflects actual model and sandbox consumption. See
[Fleet and usage](/platform/fleet-and-usage) for where those numbers originate.

## The runs list

For a focused view of run activity, `/agent/runs` lists recent runs with status
filter tabs (All, Live, Completed, Failed) and a text filter, kept current by an
interval poll plus the org-change stream. Filter to Live when you only want
what is executing right now.

:::tip
The dashboard and the runs list both refetch on the ambient org-change stream, so
starting or finishing a run updates them without a manual refresh. That stream
carries invalidations only, not the run data itself. See
[Realtime](/architecture/realtime).
:::