API

Models

List model supply visible to the workspace

GET /v1/models returns the current workspace-visible model catalogue. The catalogue is dynamic and can change as provider supply, health and policy change.

List models

Authenticate with the same workspace API key used for inference.

curl https://api.octoryn.dev/v1/models \
  -H "Authorization: Bearer $OCTORYN_API_KEY"

Response shape

The response is an object=list envelope. data contains model entries and _meta describes the current catalogue snapshot.

{
  "object": "list",
  "data": [{
    "id": "gpt-4o",
    "object": "model",
    "owned_by": "octoryn",
    "provider": "openai",
    "provider_label": "Powered by OpenAI",
    "is_byok_compatible": true,
    "context_window": 128000,
    "pricing_tier": "frontier"
  }],
  "_meta": {
    "total": 1,
    "sources": {},
    "cached_until_ms": 0,
    "expanded": []
  }
}

Per-model fields

Clients should depend only on fields they need and tolerate additional fields.

  • id — model identifier accepted in compatible inference requests
  • provider and provider_label — current supply attribution
  • is_byok_compatible — whether customer-owned credentials can serve the model
  • context_window — reported context capacity where known
  • pricing_tier — relative operating category where published
  • available_providers — present only when provider expansion is requested and available

Provider expansion

Use ?expand=providers when an operator needs the eligible provider set for each model. The expanded payload is larger and may include provider_id, model_id, status, latency_ms and kind.

GET /v1/models?expand=providers
NextModels and policy routes