---
title: "List Inference Catalog"
method: GET
path: "/api/v1/edges/inference/catalog"
tags: ["Edge"]
---

# List Inference Catalog

`GET /api/v1/edges/inference/catalog`

## Request

- Base URL: `https://example.conductor.one`
- URL: `https://example.conductor.one/api/v1/edges/inference/catalog`
- Auth: HTTP bearer and OAuth 2
  - Client credentials
  - Token URL: `/auth/v1/token`

ListInferenceCatalog returns the providers, variants and suggested models that inference routes can be built
 from, C1's registered models by family, and the model each C1 tier uses for the caller's tenant. The Edge
 feature is required.

## Response `200`

EdgeServiceListInferenceCatalogResponse contains every catalog entry.

- C1ApiEdgeV1EdgeServiceListInferenceCatalogResponse: EdgeServiceListInferenceCatalogResponse contains every catalog entry.
  - `capabilitySummary` C1ApiEdgeV1EdgeCapabilitySummary: EdgeCapabilitySummary describes the capability reports of a tenant's connected Edge processes.
    - `lastReportedAt` string, date-time, nullable
    - `reportingProcesses` integer: Edge processes with a live report.
  - `entries` C1ApiEdgeV1EdgeInferenceCatalogEntry[], nullable: The entries field.
    - `channel` 'MCP_SERVER_CATALOG_CHANNEL_UNSPECIFIED' | 'MCP_SERVER_CATALOG_CHANNEL_STABLE' | 'MCP_SERVER_CATALOG_CHANNEL_BETA' | 'MCP_SERVER_CATALOG_CHANNEL_ALPHA': Release channel; higher values are less stable.
    - `credentialKind` 'EDGE_INFERENCE_CATALOG_CREDENTIAL_KIND_UNSPECIFIED' | 'EDGE_INFERENCE_CATALOG_CREDENTIAL_KIND_API_KEY' | 'EDGE_INFERENCE_CATALOG_CREDENTIAL_KIND_AWS_WORKLOAD_IDENTITY_REGION': The credentialKind field.
    - `displayName` string: The displayName field.
    - `iconKey` string: Key the client maps to a provider icon.
    - `id` string: Entry identifier.
    - `secretContentType` string: Content type the credential secret must have for API key entries.
    - `variants` C1ApiEdgeV1EdgeInferenceCatalogVariant[], nullable: The variants field.
      - `availabilityWarning` string: Set when only some connected Edge processes can serve the variant, as during a rollout.
      - `credentialRecipe` string: Credential recipe for AWS variants; empty otherwise.
      - `credentialSource` 'EDGE_INFERENCE_CREDENTIAL_SOURCE_UNSPECIFIED' | 'EDGE_INFERENCE_CREDENTIAL_SOURCE_TENANT_VAULT' | 'EDGE_INFERENCE_CREDENTIAL_SOURCE_AWS_WORKLOAD_IDENTITY': The credentialSource field.
      - `defaultSupportedReasoningEfforts` string[], nullable: Reasoning effort levels assumed for a model id that is not suggested, lowest first. Empty means unknown.
      - `destination` string: Destination written to the target. AWS variants use a template with {region}.
      - `displayName` string: Human-readable variant name.
      - `id` string: Variant identifier, unique within its entry.
      - `modelHint` string: Placeholder describing the accepted model id shape, for the free-text field.
      - `modelIdGrammar` string: The modelIdGrammar field.
      - `nativeFormat` 'EDGE_INFERENCE_NATIVE_FORMAT_UNSPECIFIED' | 'EDGE_INFERENCE_NATIVE_FORMAT_OPENAI_RESPONSES' | 'EDGE_INFERENCE_NATIVE_FORMAT_OPENAI_CHAT' | 'EDGE_INFERENCE_NATIVE_FORMAT_ANTHROPIC_MESSAGES': Client format the variant serves.
      - `providerProfile` string: Provider profile written to the target.
      - `suggestedModels` C1ApiEdgeV1EdgeInferenceCatalogModel[], nullable: The suggestedModels field.
        - `displayName` string: Human-readable model name.
        - `id` string: Upstream model id to send.
        - `supportedReasoningEfforts` string[], nullable: Reasoning effort levels the model accepts (`none`, `minimal`, `low`, `medium`, `high`), lowest first. Empty means C1 does not know, so any level may be offered.
      - `unavailable` boolean: True when every Edge process this tenant has connected has reported that it cannot serve this variant. False, including when no Edge has reported, means the variant is offered as before. Region-scoped variants are not gated here because the region is chosen later.
      - `unavailableReason` string: Why the variant is unavailable. Set only with unavailable.
      - `upstreamPath` string: The upstreamPath field.
  - `families` C1ApiEdgeV1EdgeInferenceModelFamily[], nullable: C1's registered models grouped by family, with the ways an Edge can serve each.
    - `family` string: Family name; models with no family are grouped under their own display name.
    - `models` C1ApiEdgeV1EdgeInferenceRegistryModel[], nullable: The models field.
      - `contextWindowTokens` integer: Zero when unknown.
      - `displayName` string: The displayName field.
      - `modelId` string: Registry model id, written to EdgeInferenceRoute.registry_model_id.
      - `routeId` string: Suggested route id: the model's name in route id form.
      - `serveOptions` C1ApiEdgeV1EdgeInferenceServeOption[], nullable: Ways to serve the model, in C1's preferred order.
        - `catalogEntryId` string: Catalog entry (provider) id.
        - `catalogVariantId` string: Catalog variant id within the entry.
        - `nativeFormat` 'EDGE_INFERENCE_NATIVE_FORMAT_UNSPECIFIED' | 'EDGE_INFERENCE_NATIVE_FORMAT_OPENAI_RESPONSES' | 'EDGE_INFERENCE_NATIVE_FORMAT_OPENAI_CHAT' | 'EDGE_INFERENCE_NATIVE_FORMAT_ANTHROPIC_MESSAGES': Client format the option serves.
        - `region` string: AWS region the target uses; empty for API-key providers.
        - `upstreamModelId` string: Upstream model id to send.
  - `tiers` C1ApiEdgeV1EdgeInferenceTierOption[], nullable: The model each C1 tier currently uses for the caller's tenant. A tier with no model is left out.
    - `aliasRouteId` string: Route id published as the model alias.
    - `modelDisplayName` string: The modelDisplayName field.
    - `modelId` string: Registry model id the tier uses.
    - `routeId` string: Route id published for the tier, e.g. c1-fast.
    - `tier` 'EDGE_INFERENCE_TIER_UNSPECIFIED' | 'EDGE_INFERENCE_TIER_FAST' | 'EDGE_INFERENCE_TIER_BALANCED' | 'EDGE_INFERENCE_TIER_BEST': The tier field.

## Changes

- **2026-10-20** `fe943bedf41b`: 8 info
  - added the optional property `capabilitySummary` to the response with the `200` status
  - added the optional property `entries/items/variants/items/availabilityWarning` to the response with the `200` status
  - added the optional property `entries/items/variants/items/defaultSupportedReasoningEfforts` to the response with the `200` status
  - added the optional property `entries/items/variants/items/suggestedModels/items/supportedReasoningEfforts` to the response with the `200` status
  - …4 more
- **2026-10-06** `918a4e749f56`: 1 info
  - endpoint added

[Operation history](https://skmtc.dev/conductorone/apis/c1-api/changes/api/v1/edges/inference/catalog/get.md)

---

[API](https://skmtc.dev/conductorone/apis/c1-api.md) - [All operations](https://skmtc.dev/conductorone/apis/c1-api/llms.txt) - [OpenAPI document](https://skmtc.dev/conductorone/apis/c1-api/revisions/fe943bedf41b?raw)