Heya, I'm trying to create an audio visualizer for an assistant created via the VAPI Web SDK. But it seems like I can't access the audio stream of the assistant in a consistent way. I tried using the [call listen](https://docs.vapi.ai/calls/call-features) feature. But the latency is far too high to be used for a realtime visualizer. Does the VAPI Web SDK even expose the audio stream?

Vapi Web SDK doesn’t yet expose the assistant’s audio stream with low enough latency. Use the Web Audio API to visualize the user’s mic input in real time, assistant audio, consider upvoting the feature request or implementing a delayed post-call visualization using recorded audio..

Thanks!

Marking this ticket as Solved ✅