hi, do the realtime models from open ai or live models from google work for someone? for me they are just not saying anything.

Configure VAD correctly, send session.update with tools, and use capture_participant_transcription to ensure the model receives and responds to input..

i am having the same issue with the ui, are there any configuration that are needed besides the realtime model and temperature >0.6?

no idea, im also trying it through the ui. is there a way to adjust VAD... in the ui?