Worked on the Coze JS project to enhance speech synthesis capabilities by introducing voice emotion support within the API. Focused on API development using TypeScript and full stack principles, the work involved adding emotion and emotion scale properties to the speech request interface, as well as updating the voice and audio output interfaces to accommodate emotional configurations. This feature enables more expressive speech output, laying the groundwork for future enhancements in the repository. The implementation demonstrates a methodical approach to extending the API surface, emphasizing maintainability and scalability while addressing the need for richer, emotionally nuanced voice synthesis in application development.