status

Print the agent status — pairing, services, runtimes, models — in human or JSON form.

ownllm status
ownllm status --json

Connects to the local IPC socket, queries the running daemon, and prints a compact summary. If the daemon isn't running, falls back to a "best effort" snapshot from disk (paired ✅/❌, last heartbeat, known runtimes) and exits with code 7 (not paired) or 1 (other).

Human output

● online    tenant=acme-prod  agent=v0.4.2
├─ tunnel       cloudflared 2025.10.1     1.2 KB/s in / 4.5 KB/s out
├─ gateway      :11435                    queue=0  last 350ms
├─ ollama       0.6.4 (mlx, metal)        2 models loaded
└─ heartbeat    ok (last 12s ago)

A single status line opens with ● (green / yellow / red) so you can grep ● ownllm status in a script.

JSON output

--json prints a stable schema — the same one heartbeats post to /api/v1/agents/heartbeat:

{
  "tenant_id": "tnt_xxx",
  "agent_version": "0.4.2",
  "status": "online",
  "runtime_capabilities": [
    {
      "kind": "ollama",
      "version": "0.6.4",
      "platform": "darwin-arm64",
      "accelerators": ["metal", "mlx"],
      "status": "ready"
    }
  ],
  "models": [
    { "id": "llama-3.3-70b", "status": "ready", "size_mb": 40000 }
  ],
  "load": { "queue_depth": 0, "last_inference_ms": 350 }
}

Exit codes

Code	Condition
`0`	Daemon is running and healthy
`1`	Daemon is running but degraded (one or more components red)
`7`	Not paired

Use the exit code for monitoring, not just the textual status.

Human output

JSON output

Exit codes

See also

On this page