CLI Reference

This page covers the two operator-facing CLIs: sie-server (Python) and sie-gateway (Rust). The sie-config and sie-mcp packages ship their own console scripts, and the Rust sie-server-sidecar runs inside worker pods. sie-server uses typer for argument parsing; sie-gateway is a Rust binary using clap.

sie-server

The inference server. Start with sie-server serve.

serve

sie-server serve [OPTIONS]

Start the SIE inference server.

Option	Default	Description
--port, -p	8080	Port to listen on
--host	0.0.0.0	Host to bind to
--device, -d	auto	Device for inference: auto (detect GPU), cuda, mps, cpu
--models-dir	The packaged models directory	Models config directory (local path, s3://, gs://, abfs://, or abfss://)
--bundle, -b	None	Bundle name to load from bundles/ dir (e.g., default)
--models, -m	None	Comma-separated model names to load (mutually exclusive with --bundle)
--preload	None	Comma-separated model names to load eagerly at startup
--local-cache	HF_HOME	Local cache directory for model weights
--cluster-cache	None	Cluster cache URL for model weights (s3://, gs://, abfs://, or abfss://)
--hf-fallback/--no-hf-fallback	true	Allow or block HuggingFace Hub downloads after cache miss
--instrumentation, -i	false	Enable batch instrumentation logging
--reload	false	Enable auto-reload for development (uses uvicorn reload)
--tracing	false	Enable OpenTelemetry tracing (exports to localhost:4317)
--log-level, -l	info	Log level (DEBUG, INFO, WARNING, ERROR); env var SIE_LOG_LEVEL
--verbose, -v	false	Enable verbose logging
--json-logs	false	Enable structured JSON logging (for Loki compatibility)

Examples:

# Start with defaults (auto-detect GPU, port 8080)
sie-server serve

# Specific port and device
sie-server serve --port 8081 --device cuda

# Load specific bundle
sie-server serve --bundle default

# Load specific models only
sie-server serve --models BAAI/bge-m3,BAAI/bge-reranker-v2-m3

# Use cloud model configs
sie-server serve --models-dir s3://my-bucket/sie-models/

# Development mode with auto-reload
sie-server serve --reload --verbose

resolve-deps

sie-server resolve-deps [OPTIONS]

Resolve and print dependencies for a bundle or model list. Used by deployment scripts.

Option	Description
--bundle, -b	Bundle name to resolve deps for
--models, -m	Comma-separated model names
--models-dir	Models directory
--cpu	Exclude CUDA-only dependencies (e.g., flash-attn) for CPU-only images
--json	Output as JSON

openapi

sie-server openapi [--output PATH] [--indent N]

Export the published OpenAPI spec as JSON. Writes to stdout by default; --output/-o writes to a file.

---

sie-gateway

Stateless Rust request gateway for elastic cloud deployments.

serve

sie-gateway serve [OPTIONS]

Start the SIE Gateway server.

Option	Default	Description
--port, -p	8080	Port to listen on
--host	0.0.0.0	Host to bind to
--worker, -w	None	Local diagnostic worker URLs for ws health checks
--kubernetes	false	Local diagnostic Kubernetes endpoint discovery for ws health checks
--k8s-namespace	default	Namespace used by local diagnostic discovery
--k8s-service	sie-worker	Service name used by local diagnostic discovery
--k8s-port	8080	Worker port used by local diagnostic discovery
--log-level, -l	info	Log level: debug, info, warning, error
--json-logs	false	Enable structured JSON logging (for Loki compatibility)
--health-mode	ws raw CLI, nats via Helm	Gateway health source (nats for SIE server sidecar heartbeats, ws for local WebSocket health)
--bundles-dir	None	Optional filesystem seed for bundle configs
--models-dir	None	Optional filesystem seed for model configs

The Helm chart renders NATS health for the SIE server sidecar path. ws mode and static worker URLs are local diagnostics for the Python /ws/status stream.

Examples:

# Production health source, normally rendered by Helm
sie-gateway serve --health-mode nats

# Local WebSocket health diagnostic
sie-gateway serve -w http://localhost:8080 --health-mode ws

openapi

sie-gateway openapi [-o PATH]

Export the gateway OpenAPI spec as JSON. Writes to stdout by default; -o/--output writes to a file.

version

sie-gateway version

Show version information.

---

Environment Variables

Many CLI options can be set via environment variables. CLI arguments override environment variables, which override defaults.

Server (sie-server):

Variable	CLI Equivalent	Description
SIE_DEVICE	--device	Inference device (cuda, mps, cpu)
SIE_MODELS_DIR	--models-dir	Models config directory
SIE_MODEL_FILTER	--models	Comma-separated model names to load
SIE_LOCAL_CACHE	--local-cache	Local cache directory for weights
SIE_CLUSTER_CACHE	--cluster-cache	Cluster cache URL (s3://, gs://, abfs://, or abfss://)
SIE_HF_FALLBACK	--hf-fallback	Allow HF Hub downloads after cache miss (true/false)
SIE_LOG_LEVEL	--log-level	Log level (DEBUG, INFO, WARNING, ERROR)
SIE_LOG_JSON	--json-logs	Enable JSON logging (true/false)
SIE_TRACING_ENABLED	--tracing	Enable OpenTelemetry tracing
SIE_GPU_TYPE	-	Override detected GPU type
SIE_MEMORY_PRESSURE_THRESHOLD_PERCENT	-	GPU memory pressure threshold (50-99)
SIE_IMAGE_WORKERS	-	Image preprocessing worker count (default: 4)
SIE_INSTRUMENTATION	-	Enable detailed instrumentation

Gateway (sie-gateway):

Variable	CLI Equivalent	Description
SIE_GATEWAY_WORKERS	--worker	Local diagnostic worker URLs for ws health checks
SIE_GATEWAY_KUBERNETES	--kubernetes	Enable local diagnostic K8s discovery (true/false)
SIE_GATEWAY_K8S_NAMESPACE	--k8s-namespace	K8s namespace for local diagnostic discovery
SIE_GATEWAY_K8S_SERVICE	--k8s-service	K8s service name for local diagnostic discovery
SIE_GATEWAY_K8S_PORT	--k8s-port	Worker port for local diagnostic discovery
SIE_GATEWAY_HEALTH_MODE	--health-mode	Gateway health source (nats for the Helm SIE server sidecar path, ws for local WebSocket runs)
SIE_NATS_URL	-	NATS URL for queued inference, SIE server sidecar health, result inboxes, and config deltas
SIE_GATEWAY_ENABLE_POOLS	-	Enable resource pools (true/false)
SIE_GATEWAY_CONFIGURED_GPUS	-	Comma-separated configured GPU types
SIE_CONFIG_SERVICE_URL	-	Config service URL for bootstrap and drift polling
SIE_LOG_JSON	--json-logs	Enable JSON logging

See Configuration for the complete list.