Command Line

Use se3labs to sign in, run a robot policy on a station, and inspect evaluation progress and results.

Replace example-station-id and example-station-addr in the commands below with your station ID and relay host:port.

se3labs --help
se3labs --version
se3labs eval run --help

se3labs auth

Command	Purpose
auth login [--no-browser]	Approve a terminal through the portal
auth whoami	Show the saved account and organizations
auth logout	Remove and revoke the saved login

See authentication for the login flow and saved credentials.

se3labs eval run

se3labs eval run \
  --station example-station-id --address example-station-addr \
  --request my_policy:REQUEST --output results.json

Option	Meaning
--station, -s	Station ID; also read from SE3LABS_STATION
--address, -a	Relay or station host:port; also read from SE3LABS_ADDRESS
--request, -r	Required module:ATTRIBUTE pointing to a JobRequest
--output, -o	Write the session outcome and episode results to JSON

The request module must be importable from the current working directory or installed in your environment. The command runs the whole session and prints the returned results.

Exit code 1 indicates recorded policy errors. Handled authentication, session rejection, connection timeout, and RPC failures exit 2. A normal return without policy errors exits 0; this does not by itself mean the job or every episode succeeded. Inspect the returned state and score sheets.

se3labs eval status

se3labs eval status \
  --station example-station-id --address example-station-addr --json

Accepts the same --station and --address options as eval run. --timeout sets the RPC deadline in seconds (default 10.0); --json prints the status protobuf as JSON with its original field names.

The relay returns the latest status it received from the station. A station with no reported status answers NOT_FOUND.