Course is currently unusable

Short Course Q&A Building Live Voice Agents with Google’s ADK
1

The second content module “Running Voice Agents on DeepLearning.AI” has no context and lacks sufficient instruction to be useful on its own. Nonetheless one can install google-adk into a local Python environment and use it if they wish. But the material should be clarified.

The second module “Build your first agent” is a notebook that seems to preclude the requirement to set this up locally. Okay, great. But on attempting to chat with an agent via the ADK UI, an error is shown:

```
{"error": "404 NOT_FOUND. {'error': {'code': 404, 'message': 'models/gemini-2.0-flash-live-001 is not found for API version v1beta, or is not supported for generateContent. Call ListModels to see the list of available models and their supported methods.', 'status': 'NOT_FOUND'}}"}
```

Am I doing something wrong here or is this material simply not ready for prime time?

1 Like
2

Answer: both.

  1. Agent interactions work via voice by clicking the microphone icon and talking.
  2. They do not work (and repro the error I posted above) if you type to the agent, e,g. writing “Hello”.

This could use a note in the docs, I think. While it is a voice agent class, it wasn’t obvious to me that interaction via text wouldn’t work, and the error is opaque.

Alternatively perhaps the ability to type can be shut off in the ADK UI to constrain interaction to voice only.

Mods please close?

1 Like
3

Voice responses are working for me, but MD file and wav file is not being generated.

1 Like
4

Same here. Not working with same error

5

The issue is related to the selection of the model. To sum up:

  • gemini-2.0-flash-live-001 should be selected for live voice interaction - please see notebook L1:Build your first agent in sections app_01 and app_02

  • gemini-2.5-flash should be selected for text (chat) interaction -please see notebook L1:Build your first agent in section app_03

Hoping this might help you

1 Like
6

Thank you @Michela_Agostini for the model info and aclaration.

Also, I’d like to add about the reading item in the course that is meant to give a starting point about the folder handling and how to test the voice agents :robot:

Can you @joncooper and @rustuyucel please try again with this guidance? or is there another specific/new issue you might be facing now? I’ll be happy to help!

1 Like