[ Across the fleet ]
Every site. One command.
One control plane for every node. Deploy, monitor, and manage your entire fleet from a single interface. No SSH. No spreadsheets.
Nodes
128
Serving
124
Fleet Uptime
99.98%
OOM Events
0
Active Deployments
ground-station-08
Svalbard, NorwayVRAM
43.2/48
tok/s
26.4
Uptime
64d
ops-center-03
Edwards AFB, CAVRAM
71.8/80
tok/s
586.2
Uptime
31d
rig-platform-11
North Sea, OffshoreVRAM
9.6/24
fps
180.4
Uptime
112d
datacenter-sg-02
Singapore, APAC WarmingVRAM
--
tok/s
--
Uptime
0s
Activity
[ Capabilities ]
Command at fleet scale.
Fig 3.1
Live fleet view
Every node, every workload, every region. GPU, memory, throughput and latency in one live view.
Fig 3.2
Pin, promote, roll back
Pin the workload each site runs. Promote it across the fleet and roll back to the last pin from one place. No SSH into individual machines.
Fig 3.3
Identity and audit
Admin, operator and viewer roles. Built to plug into your identity provider. Prompt and response content is never captured.
Fig 3.4
Benchmarks and reporting
Throughput, latency, and cost measured across the fleet. Exportable scorecards for your security team or procurement review.
[ Architecture ]
Outbound only. Always.
Every node opens the connection. The control plane never reaches into your network. Works behind NAT, on dynamic IPs, and in environments where inbound is impossible.
Fig 4.1 · Outbound topology
Outbound only
The node opens the connection. The control plane never reaches in. Works through NAT and dynamic IPs.
Survives disconnect
If the link drops, the node keeps running its pinned workload. Status resyncs when the link returns.
Take command of the fleet.
Self-hosted or managed. Roles, audit, rollbacks, and live telemetry on every node.