Add OrcaRouter as a built-in model provider

[clementguarino06510-glitch]

Is your feature request related to a problem? Please describe.

Data Designer ships a set of built-in model providers (nvidia, openai, openrouter) plus pre-configured model aliases for each. OrcaRouter — an OpenAI-compatible AI gateway that exposes a provider/model namespace across many upstream models — is currently missing, so users who want adaptive routing / automatic failover / zero-markup inference through a single endpoint must hand-roll a custom provider with a base URL instead of using a first-class built-in.

Describe the solution you'd like

Add orcarouter as a built-in model provider registered in PREDEFINED_PROVIDERS and PREDEFINED_PROVIDERS_MODEL_MAP, mirroring the existing openrouter wiring:

• Provider: name=orcarouter, endpoint=https://api.orcarouter.ai/v1, provider_type=openai, api_key=ORCAROUTER_API_KEY
• Model aliases: orcarouter-text, orcarouter-reasoning, orcarouter-vision, orcarouter-embedding
• Update the default-model-settings fern docs and README quick-start so the new provider is discoverable.

Describe alternatives you've considered

Using a custom ModelProvider with an arbitrary base URL. This works, but it treats OrcaRouter as an anonymous endpoint and does not surface it as a first-class provider in data-designer config or the docs.

Agent Investigation

I explored the codebase to confirm the wiring pattern:

• packages/data-designer-config/src/data_designer/config/utils/constants.py defines OPENROUTER_PROVIDER_NAME, OPENROUTER_API_KEY_ENV_VAR_NAME, PREDEFINED_PROVIDERS, and PREDEFINED_PROVIDERS_MODEL_MAP. A new orcarouter entry slots in exactly parallel to openrouter.
• packages/data-designer-config/tests/config/test_default_model_settings.py asserts the builtin provider/model counts (3 providers, 12 model configs) and would need updating (→ 4 providers, 16 model configs).
• The ModelFacade treats providers generically via provider_type="openai"; the only provider-specific code path is OpenRouter attribution-header injection, which is optional and not required for a new OpenAI-compatible provider.
• Fern docs list the built-in providers in fern/versions/latest/pages/concepts/models/default-model-settings.mdx, index.mdx, model-providers.mdx, and custom-model-settings.mdx.

Checklist

• I've reviewed existing issues and the documentation
• This is a design proposal, not a "please build this" request