We are using the web sdk (for a chat between a user and an assistant with voice) and are streaming the responses. We're a seeing a syncing issue with the voice of the assistant. The voice data and the chat data are coming at different speeds. The chat data is available for the whole answer but the voice takes some time. How do we solve this?

Use synchronous mode async: false, optimize STT/LLM/TTS latency, implement audio caching, adjust assistant delay settings, and minimize network latency..