Models and capabilities

List models

models = client.models.list()

Native requests use teen-v1. Model IDs from patched providers are routing aliases and are rewritten before the TeenageWorks request is sent.

Retrieve capabilities

capabilities = client.capabilities.retrieve()

The response describes supported formats, streaming transports, reference and voice-description modelling, generation controls, registered LoRAs, and unavailable features.

Read capabilities during application startup or cache them for a bounded period. Do not permanently embed a copied response because server capabilities can change.

The native client raises NotFoundError for unknown model IDs. Provider patches pass unknown provider models back to the provider.