docs
CLI

Compute

See the GPU and instance capacity your org has provisioned right now

urun compute list shows the compute slices your org currently has provisioned, broken down per (app, function, compute_shape).

urun compute list
APP            FUNCTION       SHAPE     GPUS
krea-realtime  krea_runtime   b200:1    1
liveavatar     avatar         b200:4    4

What this view is — and isn't

This is what is running right now, not a catalogue or a history:

  • Slices with no provisioned GPU units are omitted server-side — you only see capacity that is actually live.
  • For the full catalogue of deployed apps, use urun app list.
  • For session history, use urun session list.

Flags

FlagDefaultEffect
--limit N100Max compute slices to fetch
--jsonoffRaw JSON instead of a table
urun compute list --json | jq '.[] | {app, function, gpus}'

Compute shape (e.g. b200:1, b200:4, h100:1) is set per function at deploy time via the @app.function decorator. To change it, edit the decorator and redeploy — it is not a runtime knob. The console's Capacity surface shows the same data with leases and QoS.

Sessions

List live and historical sessions for your org from the CLI

Trusted JWKS

Register an external JWT provider so your own users can open sessions

On this page

What this view is — and isn'tFlags