CLI Commands
Use vllm-sr to configure and run the router, manage recipes, send requests, and evaluate routing. See Installation to install the CLI and Local Docker deployment to start a router.
This reference is generated from the registered CLI commands. Command descriptions, arguments, options, and declared defaults match the source. An option without a declared default is omitted until supplied; commands may resolve it from configuration or the environment as described in their help. Repeatable options collect values. Run vllm-sr COMMAND --help for help from your installed version.
Command index
| Command | Description |
|---|---|
vllm-sr | vLLM Semantic Router CLI - Signal-driven routing across LLM providers, with a built-in model runtime. |
vllm-sr benchmark | Prepare, run, inspect, and compare sr-bench 1.0 evaluations. |
vllm-sr benchmark cancel | Cancel remaining work while retaining all existing evidence. |
vllm-sr benchmark candidate-plan | Reuse a terminal baseline's frozen protocol without repeating its requests. |
vllm-sr benchmark catalog | Show the nine benchmark adapters and evaluation profiles. |
vllm-sr benchmark compare | Compare matched cases against the best observed single model. |
vllm-sr benchmark comparison-options | List eligible baselines, or comparable live runs for BASELINE; no model calls. |
vllm-sr benchmark dataset | Prepare reproducible fixed benchmark case sets. |
vllm-sr benchmark dataset combine | Create a reusable multi-benchmark dataset from prepared manifests. |
vllm-sr benchmark dataset exclusions | Freeze finite named history without inspecting outcomes or running models. |
vllm-sr benchmark dataset options | List downloadable sources, profiles, access notes, and dependencies. |
vllm-sr benchmark dataset preparations | List shared download jobs or inspect one preparation by ID. |
vllm-sr benchmark dataset prepare | Download and freeze a dataset through the shared service by default. |
vllm-sr benchmark dataset show | Inspect a frozen dataset or list datasets in the shared store. |
vllm-sr benchmark experiment | Group durable baseline, routing checks, candidates and validation runs. |
vllm-sr benchmark experiment attach | Link existing evidence without changing or rerunning it. |
vllm-sr benchmark experiment create | Create an experiment without submitting model work. |
vllm-sr benchmark experiment delete | Delete a finished experiment's grouping and links; keep every run and result. |
vllm-sr benchmark experiment list | Read one page of experiments. |
vllm-sr benchmark experiment show | Read the experiment and one page of its linked runs. |
vllm-sr benchmark export | Export a dev response matrix for training; holdout export is rejected. |
vllm-sr benchmark plan | Validate and freeze all cases, targets, profiles, and limits without inference. |
vllm-sr benchmark preview | Inspect routing decisions without producing quality scores. |
vllm-sr benchmark reconcile-usage | Append an offline accounting correction from saved streams; no inference. |
vllm-sr benchmark recover | Create a separate attempt from a reviewed recovery plan; never auto-retry. |
vllm-sr benchmark recover-plan | Inspect eligible continuation cells without making model requests. |
vllm-sr benchmark regrade | Regrade saved MCQ/grid final outputs without mutating original evidence. |
vllm-sr benchmark replay | Estimate eligible static routes from saved answers without inference. |
vllm-sr benchmark replay-options | List eligible baselines, or compatible previews for BASELINE; no model calls. |
vllm-sr benchmark report | Show quality, four-bucket usage, cost, latency, time, and limitations. |
vllm-sr benchmark run | Execute one frozen live evaluation through the shared service. |
vllm-sr benchmark runs | |
vllm-sr benchmark serve | Own the durable journal and workers independently of a browser. |
vllm-sr benchmark setup | Inspect prerequisites or explicitly install optional benchmark harnesses. |
vllm-sr benchmark show | Read a run, bounded evidence page, or one complete saved call. |
vllm-sr benchmark target | Manage the operator-owned target registry used by the Dashboard. |
vllm-sr benchmark target list | |
vllm-sr benchmark target register | Replace the local registry from a JSON list of credential references. |
vllm-sr completion | Generate or install shell completion for vllm-sr. |
vllm-sr completion install | Install shell completions into your shell configuration. |
vllm-sr completion show | Print the completion script for a shell. |
vllm-sr config | Print generated configuration or run config subcommands. |
vllm-sr config apply | Plan, compare-and-swap, persist, and hot-reload a configuration. |
vllm-sr config envoy | Print the generated Envoy configuration. |
vllm-sr config get | Read the active canonical configuration from a Router. |
vllm-sr config init | Create a minimal canonical configuration template. |
vllm-sr config migrate | Migrate a config file to canonical v0.5 YAML. |
vllm-sr config plan | Validate and plan an exact remote mutation without changing the Router. |
vllm-sr config propose | Print a reviewable config proposal. This command does not apply it. |
vllm-sr config rollback | Compare-and-swap the active configuration to a recorded version. |
vllm-sr config router | Print the canonical router configuration. |
vllm-sr config schema | Discover the canonical config contract progressively. |
vllm-sr config validate | Validate configuration file. |
vllm-sr config versions | List the configuration history, newest first. |
vllm-sr connect | Point an agent harness at a Router gateway (System Two). |
vllm-sr connect claude-code | Route Claude Code's model calls through the gateway. |
vllm-sr connect codex | Route Codex CLI through the gateway with a profile of its own. |
vllm-sr dashboard | Open the dashboard in your default web browser. |
vllm-sr instance | Inspect the serving state of an existing local instance. |
vllm-sr instance models | Print actual native model cards for readiness checks, without inference. |
vllm-sr instance status | Print desired/observed mode and durable operation state. |
vllm-sr logs | Show logs from vLLM Semantic Router service. |
vllm-sr optimize | Analyze routing evidence and produce candidate recipe changes. |
vllm-sr optimize recipe-learning | Analyze replay and outcomes to produce recipe-learning artifacts. |
vllm-sr recipe | Validate, plan, apply, inspect, or package routing recipes. |
vllm-sr recipe apply | Validate then compare-and-swap one recipe into the active config. |
vllm-sr recipe builtin | Discover, export, or bind installed Recipes without a running Router. |
vllm-sr recipe builtin export | Export an exact five-file BUNDLE to a new directory. |
vllm-sr recipe builtin init | Bind one installed recipe to explicit models and validate the new config. |
vllm-sr recipe builtin list | List installed bundles, recipe names, and required candidate pools as JSON. |
vllm-sr recipe delete | Compare-and-swap deletion of an unreferenced named recipe. |
vllm-sr recipe get | Read one managed recipe. |
vllm-sr recipe list | List recipes and the collection ETag. |
vllm-sr recipe pack | Create a deterministic ZIP from an exact five-file RECIPE_DIR. |
vllm-sr recipe plan | Validate a recipe and bind the plan to the current config ETag. |
vllm-sr recipe validate | Validate a recipe against the running Router without changing config. |
vllm-sr request | Send requests through the stack's listener. |
vllm-sr request chat | Send a one-shot chat completion through the Envoy-routed HTTP API. |
vllm-sr route | Preview routing decisions or probe the routed inference path. |
vllm-sr route preview | Preview signals and model selection without generating an answer. |
vllm-sr route probe | Probe a real route and assert complete assistant delivery for expected 2xx. |
vllm-sr serve | Start vLLM Semantic Router. |
vllm-sr status | Show status of vLLM Semantic Router services. |
vllm-sr stop | Stop vLLM Semantic Router. |
vllm-sr storage | Inspect Router storage or manage local storage credentials. |
vllm-sr storage rotate | Replace this stack's storage credentials with freshly generated ones. |
vllm-sr storage vector-stores | List vector stores known to the Router management API. |
vllm-sr
Usage: vllm-sr [OPTIONS] [COMMAND] [ARGS]...
vLLM Semantic Router CLI - Signal-driven routing across LLM providers, with a built-in model runtime.
| Parameter | Description |
|---|---|
--version | Show version and exit. Default: false. |
--help | Show this message and exit. Default: false. |
vllm-sr benchmark
Usage: vllm-sr benchmark [OPTIONS] COMMAND [ARGS]...
Prepare, run, inspect, and compare sr-bench 1.0 evaluations.
| Parameter | Description |
|---|---|
--url TEXT | Shared sr-bench service URL; discovers the current local stack by default. Environment: SR_BENCH_URL. |
--store PATH | Durable service store; discovers the current local stack by default. Environment: SR_BENCH_STORE. |
--no-autostart | Require an already running service. Default: false. |
--help | Show this message and exit. Default: false. |
vllm-sr benchmark cancel
Usage: vllm-sr benchmark cancel [OPTIONS] RUN_ID
Cancel remaining work while retaining all existing evidence.
| Parameter | Description |
|---|---|
RUN_ID | Required argument. Type: text. |
--help | Show this message and exit. Default: false. |
vllm-sr benchmark candidate-plan
Usage: vllm-sr benchmark candidate-plan [OPTIONS] BASELINE_RUN_ID
Reuse a terminal baseline's frozen protocol without repeating its requests.
Failed, cancelled and interrupted full-plan baselines may supply the same questions and settings. This does not qualify their measurements for Compare.
| Parameter | Description |
|---|---|
BASELINE_RUN_ID | Required argument. Type: text. |
--target TEXT | Registered MoM target; may be repeated. [required] May be repeated. |
--mode CHOICE | [default: live] Choices: live, preview. |
--name TEXT | — |
--experiment TEXT | — |
--hypothesis TEXT | Default: . |
--help | Show this message and exit. Default: false. |
vllm-sr benchmark catalog
Usage: vllm-sr benchmark catalog [OPTIONS]
Show the nine benchmark adapters and evaluation profiles.
| Parameter | Description |
|---|---|
--help | Show this message and exit. Default: false. |
vllm-sr benchmark compare
Usage: vllm-sr benchmark compare [OPTIONS] BASELINE_RUN_ID CANDIDATE_RUN_ID
Compare matched cases against the best observed single model.
| Parameter | Description |
|---|---|
BASELINE_RUN_ID | Required argument. Type: text. |
CANDIDATE_RUN_ID | Required argument. Type: text. |
--help | Show this message and exit. Default: false. |
vllm-sr benchmark comparison-options
Usage: vllm-sr benchmark comparison-options [OPTIONS] [BASELINE]
List eligible baselines, or comparable live runs for BASELINE; no model calls.
| Parameter | Description |
|---|---|
[BASELINE] | Optional argument. Type: text. |
--after TEXT | Opaque cursor from the previous eligible options page. |
--limit INTEGER RANGE | [default: 10; 1<=x<=25] |
--help | Show this message and exit. Default: false. |