Introducing next-generation audio models in the API

For the first time, builders can even instruct the text-to-speech mannequin to talk in a particular method—for instance, “speak like a sympathetic customer support agent”—unlocking a brand new degree of customization for voice brokers.