Command-line reference

Overview

Commands and options for running agent samples and their model-server stacks. Worker and service entry points managed by those launchers are outside this catalog.

Sample commands

This reference is generated from the installed commands in each top-level sample’s pyproject.toml under agent-samples/, and the literal argparse declarations in its entry-point module.

conversation_example_model_servers

Launch the local model servers required by conversation-example.

Run from agent-samples/conversation-example/model-servers/:

$ uv run conversation_example_model_servers [--stop] [--gpu-profile NAME]

Options

--stop

Stop this sample’s configured model servers and exit.

--gpu-profile NAME

Select a startup hardware profile; with –stop, apply its port overrides without GPU detection.

conversation_example

Run the conversation-example hub and worker processes.

Run from agent-samples/conversation-example/:

$ uv run conversation_example [--session-recording]

Options

--session-recording

enable session recording of participant video, bidirectional audio, and data-channel traffic

lab_instrument_monitoring_model_servers

Launch the local model servers required by lab-instrument-monitoring.

Run from agent-samples/lab-instrument-monitoring/model-servers/:

$ uv run lab_instrument_monitoring_model_servers [--stop] [--gpu-profile NAME]

Options

--stop

Stop this sample’s configured model servers and exit.

--gpu-profile NAME

Select a startup hardware profile; with –stop, apply its port overrides without GPU detection.

lab_instrument_monitoring

Launch lab instrument monitoring with reused model dependencies.

Run from agent-samples/lab-instrument-monitoring/:

$ uv run lab_instrument_monitoring [--expose-web-events] [--session-recording] [--hub-config PATH] [--health-port PORT]

Options

--expose-web-events

bind the unauthenticated event viewer to all IPv4 interfaces instead of loopback

--session-recording

enable session recording of participant video, bidirectional audio, and data-channel traffic

--hub-config PATH

use hub YAML instead of the sample default (relative to launch directory)

--health-port PORT

serve stack process health on loopback PORT (1–65535; disabled by default)

procedure_capture_guide_model_servers

Launch the speech, language, and vision model servers for the sample.

Run from agent-samples/procedure-capture-guide/model-servers/:

$ uv run procedure_capture_guide_model_servers [--stop] [--gpu-profile NAME]

Options

--stop

Stop this sample’s configured model servers and exit.

--gpu-profile NAME

Select a startup hardware profile; with –stop, apply its port overrides without GPU detection.

procedure_capture_guide

Capture a demonstrated procedure or guide a participant through approved steps.

Run from agent-samples/procedure-capture-guide/:

$ uv run procedure_capture_guide [--mode {capture,guide}] [GUIDE_PATH] [--output-root PATH] [--session-recording] [--hub-config PATH] [--health-port PORT]

Positional arguments

GUIDE_PATH

approved guide YAML file or directory containing exactly one top-level YAML file; absolute or relative to the sample directory (guide mode only)

Options

--mode {capture,guide}

worker capability to run

--output-root PATH

save task packets under PATH/sessions; overrides YAML artifacts_dir (relative to the launch directory)

--session-recording

record participant video, bidirectional audio, and data-channel traffic for the whole connection in either mode, independently of procedure capture or guide commands

--hub-config PATH

use hub YAML instead of the sample default (relative to launch directory)

--health-port PORT

serve stack process health on loopback PORT (1–65535; disabled by default)

simple_vlm_example_model_servers

Launch the local model servers required by simple-vlm-example.

Run from agent-samples/simple-vlm-example/model-servers/:

$ uv run simple_vlm_example_model_servers [--stop] [--gpu-profile NAME]

Options

--stop

Stop this sample’s configured model servers and exit.

--gpu-profile NAME

Select a startup hardware profile; with –stop, apply its port overrides without GPU detection.

simple_vlm_example

simple-vlm-example orchestrator — vision Q&A over voice or text.

Run from agent-samples/simple-vlm-example/:

$ uv run simple_vlm_example [--session-recording] [--hub-config PATH] [--health-port PORT]

Options

--session-recording

enable session recording of participant video, bidirectional audio, and data-channel traffic

--hub-config PATH

use hub YAML instead of the sample default (relative to launch directory)

--health-port PORT

serve stack process health on loopback PORT (1–65535; disabled by default)

tea_making_sample_model_servers

Launch the local model servers required by tea-making-sample.

Run from agent-samples/tea-making-sample/model-servers/:

$ uv run tea_making_sample_model_servers [--stop] [--gpu-profile NAME]

Options

--stop

Stop this sample’s configured model servers and exit.

--gpu-profile NAME

Select a startup hardware profile; with –stop, apply its port overrides without GPU detection.

tea_making_sample

Launch native tea guidance with reused model services.

Run from agent-samples/tea-making-sample/:

$ uv run tea_making_sample [--expose-web-events] [--session-recording] [--hub-config PATH] [--health-port PORT]

Options

--expose-web-events

bind the unauthenticated event viewer to all IPv4 interfaces instead of loopback

--session-recording

enable session recording of participant video, bidirectional audio, and data-channel traffic

--hub-config PATH

use hub YAML instead of the sample default (relative to launch directory)

--health-port PORT

serve stack process health on loopback PORT (1–65535; disabled by default)

xr_render_demo_model_servers

Launch the local model servers required by xr-render-demo.

Run from agent-samples/xr-render-demo/model-servers/:

$ uv run xr_render_demo_model_servers [--stop] [--gpu-profile NAME]

Options

--stop

Stop this sample’s configured model servers and exit.

--gpu-profile NAME

Select a startup hardware profile; with –stop, apply its port overrides without GPU detection.

xr_render_demo

xr-render-demo orchestrator. Runs the process stack for this sample.

Run from agent-samples/xr-render-demo/:

$ uv run xr_render_demo [--session-recording] [--hub-config PATH] [--health-port PORT]

Options

--session-recording

enable session recording of participant video, bidirectional audio, and data-channel traffic

--hub-config PATH

use hub YAML instead of the sample default (relative to launch directory)

--health-port PORT

serve stack process health on loopback PORT (1–65535; disabled by default)