@langwatch/scenario
    Preparing search index...

    Variable COMPOSABLE_VOICE_LLM_MODELConst

    COMPOSABLE_VOICE_LLM_MODEL: "gpt-5.4-mini"

    Composable voice agent LLM — text-only, drives the brain of the ComposableVoiceAgent and ElevenLabsVoiceAgent stacks (STT → LLM → TTS). Python parity: COMPOSABLE_VOICE_LLM_MODEL in python/scenario/config/voice_models.py. The Python identifier is the litellm-style string "openai/gpt-5.4-mini"; the TS branded preset resolves this to openai("gpt-5.4-mini") via the @ai-sdk/openai factory.