The settings associated with speech to text for the voice assistant. This is only relevant if the assistant uses a text-to-text language model. Any assistant using a model with native audio support (e.g. fixie-ai/ultravox-v0_4) will ignore this field.
{
"model": "distil-whisper/distil-large-v2"
}