How to define the voice and language for the gemini live api while deploying a voice streaming agent?

[saurabh-emvo]

No description provided.

[hangfei]

Use the https://github.com/google/adk-python/blob/main/src/google/adk/agents/run_config.py#L41.

Good first issue: update the documentation for the use of speech config.

[kavinkumar807]

@hangfei @boyangsvl shall I update the documentation for this?

[boyangsvl]

Yes, please go ahead! Thanks!

[kavinkumar807]

@boyangsvl Correct me if I'm wrong, the documentation changes has to be done in this adk-docs repo ?

[boyangsvl]

Yes, you are right.

[kavinkumar807]

@boyangsvl I've added a PR #186 for the completeRunconfig documentation. However I couldn't able to add any full code example for audio streaming. I've tried multiple approaches even tried debugging the angular build code which uses the FastAPI endpoint to build the html code and FastAPI code equivalent to the quickstart( streaming ) documentation. But I was not able to implement it. Can I get any inputs here to add this audio streaming example?? or is it already in your backlog

[henrylawson]

Are these parameters configurable only via 5. Building a Custom Streaming App (Optional) or is there a way to set them using 4. Try the agent with adk web? Making this explicit in the docs would help users too.

[kavinkumar807]

@henrylawson As far as I've explored on ADK implementation and based on the documentation Runconfig parameters are configurable only with the run_live() as mentioned in the 5. Building a Custom Streaming App (Optional)

live_events = runner.run_live(
        session=session,
        live_request_queue=live_request_queue,
        run_config=run_config,
    )

And also adk web which uses the angular code has this audio functionality implemented which consumes this FastAPI endpoint present in the adk implementation. There is no other way to set these parameters in adk web

Cc: @boyangsvl @hangfei