Any chance Google’s own example code doesn’t work correctly?
I’m trying to test the simplest voice conversation with Gemini Live API, using the exact code from this example [ Interactive conversations with the Live API | Generative AI on Vertex AI | Google Cloud ] (Python Gen AI SDK)
And it works well for one turn - I say “Hi”, I can hear the model responding, but any subsequent voice audio is either not being sent or the response audio is not being received because the script seems to “hang” or rather the model stops responding after the first response.
I tried spinning up my own scripts using the examples but keeping running into the same issue. I tried with different models, native audio and the “legacy” ones. Issue keeps persisting.
Be on this for two days straight, would love some insight! Thanks!