Support of: gemini-2.5-flash-preview-tts [ 500 INTERNAL ]

[pretbc]

** Please make sure you read the contribution guide and file the issues in the right place. **
Contribution guide.

Describe the bug
Is gemini-2.5-flash-preview-tts supported ?

when creating:

generate_content_config = GenerateContentConfig(
    temperature=0.5,
    response_modalities=["AUDIO"],
)

root_agent = Agent(
    model="gemini-2.5-flash-preview-tts", 
    name="some name",
    description="some desc",
    instruction=agent_instruction,
    output_key="user_mood",
    generate_content_config=generate_content_config,

)

and then run run_async

class VoiceAgentCreator(AgentRunnerCreator):
    def __init__(self, session_id: str, agent: Agent, app_name: str):
        super().__init__(session_id, agent, app_name)

    async def create(self) -> None:
        await self.init_session()
        self.init_runner()
        self.init_speech_config()
        self.init_run_config()

    
    def init_speech_config(self) -> None:
        self.speech_config = types.SpeechConfig(
            voice_config=types.VoiceConfig(
                prebuilt_voice_config=types.PrebuiltVoiceConfig(voice_name="Puck")
            ),
            language_code="pl-PL"
        )

    def init_run_config(self):
        self.run_config = RunConfig(
            speech_config=self.speech_config
            )


------ then in the code after init I call:


async def call_voice_agent(voice_agent: VoiceAgentCreator, text: str):
            async for _event in voice_agent.runner.run_async(
                user_id=voice_agent.session.user_id,
                session_id=voice_agent.session.id,
                new_message=types.Content(role="user", parts=[types.Part.from_text(text=text)]),
                run_config=voice_agent.run_config,
            ):
                        
                if _event.content and _event.content.parts:
                    _part = _event.content.parts[0]
                    audio_data = _part.inline_data and _part.inline_data.data
                    if audio_data:
                        return {
                            "mime_type": "audio/pcm",
                            "data": base64.b64encode(audio_data).decode("ascii"),
                            "role": "model",
                        }

which should send message via websocket to by played on website

                message = await call_voice_agent(voice_agent, part.text)
                await websocket.send_text(json.dumps(message))

but I got error 500 from server

To Reproduce
Steps to reproduce the behavior:

1. use gemini-2.5-flash-preview-tts as agent model
2. put query to the model and send it via websocket

Expected behavior
On website I should hear agent response

Desktop (please complete the following information):

• OS: Macbook
• Python version(python -V): 3.13
• ADK version(pip show google-adk): 1.1.1

Additional context

ERROR:    Exception in ASGI application
Traceback (most recent call last):
  File "/Users/pretbc/Workspace/mood-mate/.venv/lib/python3.13/site-packages/uvicorn/protocols/websockets/websockets_impl.py", line 244, in run_asgi
    result = await self.app(self.scope, self.asgi_receive, self.asgi_send)  # type: ignore[func-returns-value]
             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/pretbc/Workspace/mood-mate/.venv/lib/python3.13/site-packages/uvicorn/middleware/proxy_headers.py", line 60, in __call__
    return await self.app(scope, receive, send)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/pretbc/Workspace/mood-mate/.venv/lib/python3.13/site-packages/fastapi/applications.py", line 1054, in __call__
    await super().__call__(scope, receive, send)
  File "/Users/pretbc/Workspace/mood-mate/.venv/lib/python3.13/site-packages/starlette/applications.py", line 112, in __call__
    await self.middleware_stack(scope, receive, send)
  File "/Users/pretbc/Workspace/mood-mate/.venv/lib/python3.13/site-packages/starlette/middleware/errors.py", line 152, in __call__
    await self.app(scope, receive, send)
  File "/Users/pretbc/Workspace/mood-mate/.venv/lib/python3.13/site-packages/starlette/middleware/exceptions.py", line 62, in __call__
    await wrap_app_handling_exceptions(self.app, conn)(scope, receive, send)
  File "/Users/pretbc/Workspace/mood-mate/.venv/lib/python3.13/site-packages/starlette/_exception_handler.py", line 53, in wrapped_app
    raise exc
  File "/Users/pretbc/Workspace/mood-mate/.venv/lib/python3.13/site-packages/starlette/_exception_handler.py", line 42, in wrapped_app
    await app(scope, receive, sender)
  File "/Users/pretbc/Workspace/mood-mate/.venv/lib/python3.13/site-packages/starlette/routing.py", line 714, in __call__
    await self.middleware_stack(scope, receive, send)
  File "/Users/pretbc/Workspace/mood-mate/.venv/lib/python3.13/site-packages/starlette/routing.py", line 734, in app
    await route.handle(scope, receive, send)
  File "/Users/pretbc/Workspace/mood-mate/.venv/lib/python3.13/site-packages/starlette/routing.py", line 362, in handle
    await self.app(scope, receive, send)
  File "/Users/pretbc/Workspace/mood-mate/.venv/lib/python3.13/site-packages/starlette/routing.py", line 95, in app
    await wrap_app_handling_exceptions(app, session)(scope, receive, send)
  File "/Users/pretbc/Workspace/mood-mate/.venv/lib/python3.13/site-packages/starlette/_exception_handler.py", line 53, in wrapped_app
    raise exc
  File "/Users/pretbc/Workspace/mood-mate/.venv/lib/python3.13/site-packages/starlette/_exception_handler.py", line 42, in wrapped_app
    await app(scope, receive, sender)
  File "/Users/pretbc/Workspace/mood-mate/.venv/lib/python3.13/site-packages/starlette/routing.py", line 93, in app
    await func(session)
  File "/Users/pretbc/Workspace/mood-mate/.venv/lib/python3.13/site-packages/fastapi/routing.py", line 383, in app
    await dependant.call(**solved_result.values)
  File "/Users/pretbc/Workspace/mood-mate/app/main.py", line 294, in websocket_endpoint
    await asyncio.gather(agent_to_client_task, client_to_agent_task)
  File "/Users/pretbc/Workspace/mood-mate/app/main.py", line 195, in agent_to_client_messaging
    message = await call_voice_agent(voice_agent, part.text)
              ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/pretbc/Workspace/mood-mate/app/main.py", line 201, in call_voice_agent
    async for _event in voice_agent.runner.run_async(
    ...<16 lines>...
                }
  File "/Users/pretbc/Workspace/mood-mate/.venv/lib/python3.13/site-packages/google/adk/runners.py", line 197, in run_async
    async for event in invocation_context.agent.run_async(invocation_context):
    ...<2 lines>...
      yield event
  File "/Users/pretbc/Workspace/mood-mate/.venv/lib/python3.13/site-packages/google/adk/agents/base_agent.py", line 147, in run_async
    async for event in self._run_async_impl(ctx):
      yield event
  File "/Users/pretbc/Workspace/mood-mate/.venv/lib/python3.13/site-packages/google/adk/agents/llm_agent.py", line 278, in _run_async_impl
    async for event in self._llm_flow.run_async(ctx):
      self.__maybe_save_output_to_state(event)
      yield event
  File "/Users/pretbc/Workspace/mood-mate/.venv/lib/python3.13/site-packages/google/adk/flows/llm_flows/base_llm_flow.py", line 279, in run_async
    async for event in self._run_one_step_async(invocation_context):
      last_event = event
      yield event
  File "/Users/pretbc/Workspace/mood-mate/.venv/lib/python3.13/site-packages/google/adk/flows/llm_flows/base_llm_flow.py", line 305, in _run_one_step_async
    async for llm_response in self._call_llm_async(
    ...<8 lines>...
        yield event
  File "/Users/pretbc/Workspace/mood-mate/.venv/lib/python3.13/site-packages/google/adk/flows/llm_flows/base_llm_flow.py", line 520, in _call_llm_async
    async for llm_response in llm.generate_content_async(
    ...<16 lines>...
      yield llm_response
  File "/Users/pretbc/Workspace/mood-mate/.venv/lib/python3.13/site-packages/google/adk/models/google_llm.py", line 146, in generate_content_async
    response = await self.api_client.aio.models.generate_content(
               ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
    ...<3 lines>...
    )
    ^
  File "/Users/pretbc/Workspace/mood-mate/.venv/lib/python3.13/site-packages/google/genai/models.py", line 7453, in generate_content
    response = await self._generate_content(
               ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
        model=model, contents=contents, config=parsed_config
        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
    )
    ^
  File "/Users/pretbc/Workspace/mood-mate/.venv/lib/python3.13/site-packages/google/genai/models.py", line 6447, in _generate_content
    response_dict = await self._api_client.async_request(
                    ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
        'post', path, request_dict, http_options
        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
    )
    ^
  File "/Users/pretbc/Workspace/mood-mate/.venv/lib/python3.13/site-packages/google/genai/_api_client.py", line 799, in async_request
    result = await self._async_request(http_request=http_request, stream=False)
             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/pretbc/Workspace/mood-mate/.venv/lib/python3.13/site-packages/google/genai/_api_client.py", line 743, in _async_request
    await errors.APIError.raise_for_async_response(response)
  File "/Users/pretbc/Workspace/mood-mate/.venv/lib/python3.13/site-packages/google/genai/errors.py", line 131, in raise_for_async_response
    raise ServerError(status_code, response_json, response)
google.genai.errors.ServerError: 500 INTERNAL. {'error': {'code': 500, 'message': 'An internal error has occurred. Please retry or report in https://developers.generativeai.google/guide/troubleshooting', 'status': 'INTERNAL'}}

[pretbc]

update: did some downgrades to google-adk 1.0.0 and the upgrade following :

The traceback indicates the error TypeError: int() argument must be a string, a bytes-like object or a real number, not 'dict' originates deep within Python's standard enum.py library (line 268). This happens during the initialization of httpcore._sync.http11.HTTPConnectionState, which is an enum.IntEnum. This means that when Python is trying to define a member of this integer-based enum (e.g., IDLE = 1), the expected integer value is somehow being presented as a dictionary.

The session attributes you printed (id='test', user_id='test') are correct strings. The issue is not with the arguments you're passing to voice_agent.runner.run_async in call_voice_agent, but rather with the underlying library stack, likely due to package version incompatibilities, especially given you are using Python 3.13.

Your pyproject.toml specifies google-generativeai>=0.7.0. The current version is significantly newer (around 1.18.0 as of this writing). An older version like 0.7.0 might not be fully compatible with more recent versions of httpx (which uses httpcore) or Python 3.13.
Similarly, your google-adk is pinned to rev = "v0.2.0", and newer versions are available (e.g., v1.1.1).

I recommend updating these key libraries.

Here's how you can do it:

1. Update google-generativeai:
   I'll use the uv add command to upgrade it.

uv add google-generativeai --upgrade

The google-generativeai package has been updated to version 0.8.5. This is an improvement, though significantly newer versions (e.g., 1.18.0+) are available. The limited upgrade might be due to other dependencies or existing constraints in your uv.lock file.

Next, let's update google-adk. In your pyproject.toml, it's pinned to rev = "v0.2.0". I'll update this to a more recent version, v1.1.1.

I will now edit pyproject.toml to change the revision for google-adk.

[tool.uv.sources]
google-adk = { git = "https://github.com/google/adk-python.git", rev = "v1.1.1" }

The pyproject.toml file should now reference v1.1.1 for google-adk.
Now, let's synchronize your environment with the updated pyproject.toml.

uv pip sync pyproject.toml

The google-adk package has been updated to v1.1.1 by synchronizing with your pyproject.toml.
However, the output shows that many packages, including httpx and httpcore (which are crucial for google-generativeai), were uninstalled. This might leave your environment in an inconsistent state.

To ensure all necessary direct and transitive dependencies from your pyproject.toml are correctly installed, I will now run a command to install based on the current project configuration.

uv pip install .

The uv pip install . command has reinstalled the necessary packages. Notably, google-genai is now at version 1.18.0, and crucial dependencies like httpx==0.28.1 and httpcore==1.0.9 are back. google-adk remains at 1.1.1 as intended.

With these updated and correctly installed dependencies, the original TypeError related to httpcore's enum initialization should hopefully be resolved.

and now I got

      yield llm_response
  File "/Users/pretbc/Workspace/mood-mate/.venv/lib/python3.13/site-packages/google/adk/models/google_llm.py", line 146, in generate_content_async
    response = await self.api_client.aio.models.generate_content(
               ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
    ...<3 lines>...
    )
    ^
  File "/Users/pretbc/Workspace/mood-mate/.venv/lib/python3.13/site-packages/google/genai/models.py", line 7453, in generate_content
    response = await self._generate_content(
               ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
        model=model, contents=contents, config=parsed_config
        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
    )
    ^
  File "/Users/pretbc/Workspace/mood-mate/.venv/lib/python3.13/site-packages/google/genai/models.py", line 6447, in _generate_content
    response_dict = await self._api_client.async_request(
                    ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
        'post', path, request_dict, http_options
        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
    )
    ^
  File "/Users/pretbc/Workspace/mood-mate/.venv/lib/python3.13/site-packages/google/genai/_api_client.py", line 799, in async_request
    result = await self._async_request(http_request=http_request, stream=False)
             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/pretbc/Workspace/mood-mate/.venv/lib/python3.13/site-packages/google/genai/_api_client.py", line 743, in _async_request
    await errors.APIError.raise_for_async_response(response)
  File "/Users/pretbc/Workspace/mood-mate/.venv/lib/python3.13/site-packages/google/genai/errors.py", line 129, in raise_for_async_response
    raise ClientError(status_code, response_json, response)
google.genai.errors.ClientError: 429 RESOURCE_EXHAUSTED. {'error': {'code': 429, 'message': 'You exceeded your current quota, please check your plan and billing details. For more information on this error, head to: https://ai.google.dev/gemini-api/docs/rate-limits.', 'status': 'RESOURCE_EXHAUSTED', 'details': [{'@type': 'type.googleapis.com/google.rpc.QuotaFailure', 'violations': [{'quotaMetric': 'generativelanguage.googleapis.com/generate_content_free_tier_requests', 'quotaId': 'GenerateRequestsPerDayPerProjectPerModel-FreeTier', 'quotaDimensions': {'model': 'gemini-2.5-flash-tts', 'location': 'global'}, 'quotaValue': '15'}]}, {'@type': 'type.googleapis.com/google.rpc.Help', 'links': [{'description': 'Learn more about Gemini API quotas', 'url': 'https://ai.google.dev/gemini-api/docs/rate-limits'}]}, {'@type': 'type.googleapis.com/google.rpc.RetryInfo', 'retryDelay': '49s'}]}}

[pretbc]

update:

after few tries and again

  File "/Users/pretbc/Workspace/mood-mate/.venv/lib/python3.13/site-packages/google/genai/_api_client.py", line 743, in _async_request
    await errors.APIError.raise_for_async_response(response)
  File "/Users/pretbc/Workspace/mood-mate/.venv/lib/python3.13/site-packages/google/genai/errors.py", line 131, in raise_for_async_response
    raise ServerError(status_code, response_json, response)
google.genai.errors.ServerError: 500 INTERNAL. {'error': {'code': 500, 'message': 'An internal error has occurred. Please retry or report in https://developers.generativeai.google/guide/troubleshooting', 'status': 'INTERNAL'}}
INFO:     connection closed

[jwiejulie]

Hi @pretbc !

We're currently doing some housekeeping on our issue tracker. Given the rapid pace of development and the number of updates since this was last active, there's a good chance this issue has been resolved in a newer version. To help us keep our backlog focused on current issues, we are closing this as stale.

If you're still facing this problem, could you please update to the very latest version of the library and confirm the issue still exists? If it does, please comment here, and we will be happy to reopen this issue and investigate further.

Thanks for your contribution and understanding!

[joaquinperaza]

Hey! just faced the same issue on "google-adk" version = "1.19.0":

INFO:httpx:HTTP Request: POST https://generativelanguage.googleapis.com/v1beta/models/gemini-2.5-flash-preview-tts:generateConten
t "HTTP/1.1 500 Internal Server Error"
Error en handle_audio: 500 INTERNAL. {'error': {'code': 500, 'message': 'An internal error has occurred. Please retry or report i
n https://developers.generativeai.google/guide/troubleshooting', 'status': 'INTERNAL'}}

[hotairbag]

honestly this api is just unusable. the amount of api 500 errors i get using gemini 2.5 tts is like 40% of the time! how is this even possible

[DigiJungle]

honestly this api is just unusable. the amount of api 500 errors i get using gemini 2.5 tts is like 40% of the time! how is this even possible

Exactly the same problem with gemini 2.5 flash lite preview tts...