Closed juberti closed 8 months ago
streaming latency comprises of network, tts, video creation and llm sometimes (Agents) the breakdown will be available in the client notification The latency was reduced significantly in the last few months - please lets us know what you think
/talks API response usually takes about 1 second, and often fails with a 429 error. Then, it usually takes another 2-3 seconds for the video to start playing. This is too slow for real-time applications.