> ## Documentation Index
> Fetch the complete documentation index at: https://evoke-f0bfabff.mintlify.site/llms.txt
> Use this file to discover all available pages before exploring further.

# Business Hub

> The live multi-project observability dashboard on :3008.

The **Business Hub** is RStack's company dashboard. It aggregates every run
across every project on the machine and shows them live — no database, no
telemetry leaves your machine; everything is derived from the `.rstack/` files.

## Launch it

The hub opens automatically when a session starts (Pi, the Claude Code
SessionStart hook, and Operator all launch it). Bring it up manually from any
framework:

```bash theme={null}
npx rstack-agents hub          # health-check :3008, start if down, open browser
```

Or run the server directly:

```bash theme={null}
rstack-business --port 3008 --project .
```

<Info>
  Default port is **3008** (override with `RSTACK_BUSINESS_PORT`). Disable
  auto-launch with `RSTACK_NO_BUSINESS_HUB=1`. The legacy `rstack-observer`
  command is a compatibility alias that opens the same hub.
</Info>

## Pages

<CardGroup cols={2}>
  <Card title="Studio & Studio 3D" icon="robot">
    A live agent workspace — every stage as a workstation, status as a glowing
    state. The 3D scene (`/studio3d`) lets you click any agent to see what it
    worked on, what it shipped, and why it's waiting.
  </Card>

  <Card title="Business Flex" icon="sliders">
    Active profile packs, enabled domains, budget guardrails, routed tasks,
    selected specialists, and routing proof from real `.rstack` task metadata.
  </Card>

  <Card title="Run Analytics" icon="chart-gantt">
    Wall-clock Gantt per run, average stage durations, and run-over-run trends.
  </Card>

  <Card title="Team & Presence" icon="users">
    Who is live now, a people directory (runs / approvals / guidance per
    person), and a manager rollup per project.
  </Card>

  <Card title="Projects & Runs" icon="folder-tree">
    Every session. Open a run to read its actual deliverables and evidence.
  </Card>

  <Card title="Approvals & Alerts" icon="shield-check">
    Approve or reject gates from the browser; popups when new gates block.
  </Card>

  <Card title="Workflow & Traceability" icon="diagram-project">
    The 15-stage map with live states and requirement → stage → task → evidence
    chains.
  </Card>
</CardGroup>

## Read what the agents produced

Open any run and the drawer lists its **deliverables** — `requirements.json`,
`architecture.md`, the QA report, security review, release readiness, the plan —
grouped by stage. Click one to read it in the dashboard. The **Evidence**
section lists the validation records beside them.

The artifact API is strictly sandboxed: the run is located via the known
project roots, the resolved path must stay inside the run directory, only
text artifacts are served, and there's a 512 KB cap.

## Switch scope & share links

A project → run switcher in the top bar scopes every page to the run you care
about and remembers your choice. Links are shareable: `#run=<runId>` deep-links
straight to a run (handy in Slack).

## Multi-project aggregation

The hub reads a global registry (`~/.rstack/known-projects.json`) that every
`rstack-agents init` and `hub` call updates, so one dashboard shows runs from
all your projects. Many people can point the same hub (or the same Slack bot)
at it.

<Note>
  **Cost status:** RStack now creates profile-level budget policies and
  task-level budget envelopes before delegated work starts. Real token/model
  usage is still host-dependent: if the host framework does not report usage,
  `actual_usd` remains builder-reported or zero. See
  [AI SDLC Trends & Loopholes](/reference/loopholes-roadmap).
</Note>