uip model-hub

Syntax and options for `uip model-hub`, a read/configure view over LLM Gateway routing rules — organizations, tenants, custom-routings, routing-history, and policy-previews.

uip model-hub is a read/configure view over LLM Gateway routing — which model backs a given (product, feature) pair, in which region, and who can override that default. Five command groups: organizations, tenants, custom-routings, routing-history, policy-previews. All require the OrganizationAdmin role for AI Trust Layer.

See uip llm-gateway for the discovery-side API this configuration ultimately feeds.

Concepts

• Product / feature / model. A routing rule is keyed by a --product (e.g. agenthub), a --feature within it (e.g. llm-call), and a --model identifier. tenants get returns every rule for the current tenant, grouped by product.
• Custom routing overrides the default. custom-routings create upserts a rule that redirects a (product, feature, primary-model) tuple from its default region to a different model/region pair; if the target you set matches the GitOps default, the override is removed rather than stored as a redundant rule.
• Regions. The valid geography codes are EU, CA, US, SI, JA, AU, IN, UK, CH, UAE, SK, GLOBAL — passing anything else to a region option fails client-side before any network call.
• Policy previews are read-only. policy-previews get computes what tenants get's shape would look like if a given governance policy were active — it does not apply anything.

Synthèse

uip model-hub organizations get [--tenant <name>]
uip model-hub tenants get [--tenant <name>] [--force-refresh]
uip model-hub custom-routings get --product <name> --feature <name> --primary-model <id> --current-model <id> [--primary-model-region <region>] [--current-region <region>] [--tenant <name>]
uip model-hub custom-routings create --product <name> --feature <name> --primary-model <id> --primary-model-region <region> --current-model <id> --current-region <region> [--tenant <name>]
uip model-hub routing-history list --product <name> --feature <name> --model <id> [--tenant <name>]
uip model-hub policy-previews get <policy-id> [--tenant <name>]

All verbs also accept --login-validity <minutes> (override the interactive-login token lifetime).

uip model-hub organizations get

Fetch the Model Hub organization overview — per-tenant model counts (total/active/custom/deprecating-soon) and blocked-entry totals.

Options

Long	Valeur (Value)	Description
--tenant <name>	string	Target tenant name (scopes which tenant's counts to include if the org spans several — otherwise informational).
--login-validity <minutes>	Integer	Override the interactive-login token lifetime.

Exemple

uip model-hub organizations get

Format des données (--sortie json)

{
  "Code": "ModelHubOrganizationsGet",
  "Data": {
    "tenants": [
      {
        "tenantName": "DefaultTenant",
        "tenantId": "00000000-0000-0000-0000-000000000000",
        "region": "westeurope",
        "totalModels": 42,
        "activeModels": 38,
        "customModels": 2,
        "deprecatingSoonModels": 1,
        "blockedEntries": 0
      }
    ]
  }
}

uip model-hub tenants get

Fetch the Model Hub tenant overview for the current tenant — every routing rule, grouped by product, with model/managed-by/status/BYO-onboarding state.

Options

Long	Valeur (Value)	Description
--tenant <name>	string	Target tenant name.
--force-refresh	Drapeau	Bypass any cache backing this view (including the resolved policy) and re-fetch from source. Default false.
--login-validity <minutes>	Integer	Override the interactive-login token lifetime.

Exemple

uip model-hub tenants get

Format des données (--sortie json)

{
  "Code": "ModelHubTenantsGet",
  "Data": {
    "products": [
      {
        "modelHubProduct": "agenthub",
        "rules": [
          {
            "feature": "agenthub-llm-call",
            "model": "anthropic.claude-sonnet-4-20250514-v1:0",
            "managedBy": "UiPath",
            "modelStatus": "Active",
            "isBlocked": false,
            "isFeatureOnboardedToByo": false
          }
        ]
      }
    ]
  }
}

uip model-hub custom-routings get

Look up the per-region model-availability map for a (product, feature, primary-model) tuple — use this to pick valid --current-model/--current-region values before running custom-routings create.

Options

Long	Valeur (Value)	Requis	Description
--product <name>	string	oui	Model Hub product, e.g. agenthub.
--feature <name>	string	oui	Feature within the product, e.g. llm-call.
--primary-model <id>	string	oui	Base rule's model identifier whose regional alternatives to enumerate.
--current-model <id>	string	oui	Required by server-side request validation and echoed back in the response — pass the same value as --primary-model if you have no other choice yet.
--primary-model-region <region>	EU|CA|US|SI|JA|AU|IN|UK|CH|UAE|SK|GLOBAL	non	Sent for request-shape completeness; not used for the lookup itself.
--current-region <region>	EU|CA|US|SI|JA|AU|IN|UK|CH|UAE|SK|GLOBAL	non	Sent for request-shape completeness; not used for the lookup itself.
--tenant <name>	string	non	Target tenant name.
--login-validity <minutes>	Integer	non	Override the interactive-login token lifetime.

Exemple

uip model-hub custom-routings get \
  --product agenthub --feature llm-call \
  --primary-model gpt-5-mini-2025-08-07 --current-model gpt-5-mini-2025-08-07

Format des données (--sortie json)

{
  "Code": "ModelHubCustomRoutingsGet",
  "Data": {
    "organizationId": "00000000-0000-0000-0000-000000000000",
    "tenantId": "00000000-0000-0000-0000-000000000000",
    "modelHubProduct": "agenthub",
    "feature": "llm-call",
    "primaryModel": "gpt-5-mini-2025-08-07",
    "currentModel": "gpt-5-mini-2025-08-07",
    "availableModelsInRegions": {
      "EU": [{ "model": "gpt-5-mini-2025-08-07", "isBlockedByPolicy": false }],
      "US": [{ "model": "gpt-5-mini-2025-08-07", "isBlockedByPolicy": false }]
    }
  }
}

uip model-hub custom-routings create

Upsert a custom routing rule for (product, feature, primary-model). If the requested target matches the GitOps default, the existing custom rule is removed instead. Returns the resulting tenant overview (same shape as tenants get).

Options

Long	Valeur (Value)	Requis	Description
--product <name>	string	oui	Model Hub product.
--feature <name>	string	oui	Feature within the product.
--primary-model <id>	string	oui	Base rule's model identifier being overridden.
--primary-model-region <region>	EU|CA|US|SI|JA|AU|IN|UK|CH|UAE|SK|GLOBAL	oui	Region of the base rule.
--current-model <id>	string	oui	Target model to route to (often the same as --primary-model).
--current-region <region>	EU|CA|US|SI|JA|AU|IN|UK|CH|UAE|SK|GLOBAL	oui	Target region to route to.
--tenant <name>	string	non	Target tenant name.
--login-validity <minutes>	Integer	non	Override the interactive-login token lifetime.

Invalid region values on either region flag fail client-side, listing the accepted set, before any network call.

Exemple

uip model-hub custom-routings create \
  --product agenthub --feature llm-call \
  --primary-model gpt-5-mini-2025-08-07 --primary-model-region US \
  --current-model gpt-5-mini-2025-08-07 --current-region EU

Format des données (--sortie json)

Same shape as tenants get, reflecting the rule immediately after the upsert — e.g. the affected rule's modelRouting/effectiveDefaultModelRouting will show "EU".

uip model-hub routing-history list

List routing-rule change history for a (product, feature, model) — old/new model and region per change, the actor, and the timestamp.

Options

Long	Valeur (Value)	Requis	Description
--product <name>	string	oui	Model Hub product.
--feature <name>	string	oui	Feature within the product.
--model <id>	string	oui	Model identifier whose history to retrieve.
--tenant <name>	string	non	Target tenant name.
--login-validity <minutes>	Integer	non	Override the interactive-login token lifetime.

Exemple

uip model-hub routing-history list \
  --product agenthub --feature llm-call --model gpt-5.1-2025-11-13

Format des données (--sortie json)

{
  "Code": "ModelHubRoutingHistoryList",
  "Data": [
    {
      "product": "agenthub",
      "feature": "llm-call",
      "oldModel": "gpt-5.1-2025-11-13",
      "oldRegion": "EU",
      "oldConfigIsByo": false,
      "newModel": "gpt-5.1-2025-11-13",
      "newRegion": "CA",
      "newConfigIsByo": false,
      "actor": "User",
      "organizationId": "00000000-0000-0000-0000-000000000000",
      "tenantId": "00000000-0000-0000-0000-000000000000",
      "createdAt": "2026-05-22T10:15:00.000Z"
    }
  ]
}

actor is "System" or "User". Data is [] when there's no history for the tuple, not an error.

uip model-hub policy-previews get

Preview the tenant Model Hub configuration that would result if a given governance policy were applied — same response shape as tenants get, computed against <policy-id> instead of the currently-active policy. Read-only; does not activate anything.

Arguments

Nom	Requis	Objectif
<policy-id>	oui	Governance policy ID to preview.

Options

Long	Valeur (Value)	Description
--tenant <name>	string	Target tenant name.
--login-validity <minutes>	Integer	Override the interactive-login token lifetime.

Exemple

uip model-hub policy-previews get 11111111-2222-3333-4444-555555555555

Format des données (--sortie json)

{
  "Code": "ModelHubPolicyPreviewsGet",
  "Data": {
    "products": [
      {
        "modelHubProduct": "agenthub",
        "rules": [
          {
            "feature": "llm-call",
            "model": "gpt-5-mini-2025-08-07",
            "managedBy": "UiPath",
            "modelStatus": "Active",
            "isBlocked": true,
            "isFeatureOnboardedToByo": false
          }
        ]
      }
    ]
  }
}

Comportement des erreurs

Every verb fails the same way on an auth/permission problem: Instructions: "Ensure you are logged in with 'uip login' and your account has the OrganizationAdmin role for AI Trust Layer." — check Message for the specific underlying error. Region-flag validation on custom-routings get/create fails before any network call, listing the accepted EU/CA/US/SI/JA/AU/IN/UK/CH/UAE/SK/GLOBAL set.

Voir également

• uip llm-gateway — LLM model discovery, the read side of what this configures.
• uip llm-configuration — BYO LLM connections referenced by BYO-onboarded routing rules.
• Outils (extensions)
• Options globales
• Codes de sortie