Multi Agent Audio support (with gemini-2.0-flash-exp)

[owennewo]

When is multi-agent audio going to be supported? I know the docs say 'examples coming' but I suspect it also needs back-end work.

I've got a cut down agent_team tutorial (root_agent (weather) + greeting_agent, farewelll_agent) and it hands off nicely in text mode if I 'say hi' but when in audio (clicking audio button in adk web) it does respond with speech but acts like it is talking to the root_agent only (no handoff) and does not emit any events - even when the session is flipped to 'streaming'.

I'm doing auth with ai studio:

GOOGLE_GENAI_USE_VERTEXAI="False"
GOOGLE_API_KEY=XXXXXXXX

There are a few errors in the logs - but I suspect you don't expect it to work yet. Correct?

[boyangsvl]

We will be working on it soon but we don't have an ETA at the moment. Sorry.