Recently, xAI released a major version update for its Grok Voice system, significantly enhancing the interactive experience and making a crucial step in voice assets. In this update, xAI introduced 21 new flagship voices at once, aiming to provide more personalized and distinctive choices for different application scenarios.
These newly added voices, along with the existing five voices, form a more rich audio resource library. According to the information, these 21 voices cover diverse fields such as customer service, role-playing, content explanation, commercial advertising, and educational tutoring, accurately adapting to the needs of various specific use cases. At the same time, xAI has also deeply retrained the five classic voices: Ara, Eve, Leo, Rex, and Sal. By optimizing rhythm, wording, and stress handling, these voices have made a qualitative leap in natural expression.
To allow developers to flexibly call, xAI simultaneously launched the Grok Voice Agent Builder platform. Currently, these voices are fully integrated into the real-time voice agent API and text-to-speech (TTS) API. Notably, all voices have native multilingual support, capable of smoothly handling more than 25 languages. The system also introduced refined voice control tags such as [pause], allowing developers to intervene more strictly in the expression rhythm of the voice.
Starting today, users can freely try out these new voice effects through the official Playground environment and explore their practical applications in real projects. This update marks xAI's continuous efforts in building a voice agent ecosystem, providing developers in related fields with more competitive technical capabilities.
