API reference

Models

Model routes expose the curated, provider-independent model catalog and compute availability from enabled, configured Linked Services in the selected workspace.

Model calls

Method	Path	Description
`GET`	`/api/v1/models`	Lists models, catalog metadata, pricing, and availability.
`GET`	`/api/v1/models/{model_id}`	Describes deployments, capabilities, and availability details.

List models

Parameter	Description
`provider_id`	Filters by `anthropic`, `nvidia`, `ollama`, `openai`, or `together`.
`capability`	Filters by `coding`, `function_calling`, `reasoning`, `structured_output`, `text`, or `vision`.
`status`	Filters by catalog status: `active`, `deprecated`, or `preview`.
`lab`	Filters by one or more repeated, typed lab slugs. Repeated values use OR semantics.
`sort`	Orders by `alphabetical-asc`, `alphabetical-desc`, `newest`, or `oldest`. Date sorts use `released_at`.
`offset`	Skips filtered results. The endpoint always returns at most 24 models.
`available`	Filters by effective workspace availability (`true` or `false`).
`search`	Searches name, vendor, family, and tags; 1-200 characters.
`include_deprecated`	Includes deprecated entries when no explicit `status` is supplied. Defaults to `false`.

Supported labs are anthropic, deep-cogito, deepseek, google, meta, minimax, mistral, moonshot, nvidia, openai, qwen, thinking-machines, and z.

Without status, deprecated models are excluded unless include_deprecated=true. An explicit status=deprecated selects deprecated models directly and does not require include_deprecated=true.

GET /api/v1/models lists catalog entries and their workspace availability.

GET /api/v1/models?capability=coding&available=true
X-Workspace-ID: workspace-acme

Example response (200 OK):

{
  "items": [
    {
      "id": "kimi-k2.6",
      "name": "Kimi K2.6",
      "vendor": "Moonshot AI",
      "family": "Kimi K2",
      "description": "A coding-capable reasoning model.",
      "capabilities": ["coding", "reasoning", "text"],
      "status": "active",
      "providers": ["together"],
      "available": true,
      "starting_price": {
        "input_per_million_usd": "0.6",
        "output_per_million_usd": "2.5",
        "cached_input_per_million_usd": null,
        "valid_until": null,
        "note": null
      },
      "tags": ["coding", "reasoning"]
    }
  ],
  "total": 1
}

List items include summary metadata, the model version’s released_at date, capabilities, provider IDs, availability, starting price, and tags. A describe response additionally includes every provider deployment, its endpoint and limits, pricing, replacement model, and workspace-specific availability reason.

Describe a model

GET /api/v1/models/{model_id} returns every configured deployment and its workspace availability.

GET /api/v1/models/kimi-k2.6
X-Workspace-ID: workspace-acme

Example response (200 OK):

{
  "id": "kimi-k2.6",
  "name": "Kimi K2.6",
  "vendor": "Moonshot AI",
  "family": "Kimi K2",
  "description": "A coding-capable reasoning model.",
  "capabilities": ["coding", "reasoning", "text"],
  "status": "active",
  "replacement_model_id": null,
  "tags": ["coding", "reasoning"],
  "deployments": [
    {
      "provider_id": "together",
      "provider_model_id": "moonshotai/Kimi-K2.6",
      "endpoint": "chat_completions",
      "context_window": 262144,
      "max_output_tokens": 32768,
      "pricing": {
        "input_per_million_usd": "0.6",
        "output_per_million_usd": "2.5",
        "cached_input_per_million_usd": null,
        "valid_until": null,
        "note": null
      },
      "status": "active",
      "deprecated_at": null
    }
  ],
  "availability": [
    {
      "provider_id": "together",
      "configured": true,
      "enabled": true,
      "available": true,
      "reason": null
    }
  ],
  "available": true
}