[Bug] Mistral Client: TypeError: 'async for' requires an object with __aiter__ method, got coroutine

[galaxygliese]

Hello, I got the error while I was using Mistral AI client in app.agno.com.

Description

Running the Mistral client as is works fine, but running ti from app.agno.com is causing the error.
I tried OpenAI client version, it worked.

Steps to Reproduce

List the steps needed to encounter this bug or issue.

Agent Configuration (if applicable)

Mistral AI model code:

def get_mistral_model():
    mistral_api_key = os.getenv("MISTRAL_API_KEY")
    client = Mistral(
        api_key=mistral_api_key,
    )
    model = MistralChat(
        id="mistral-large-latest",
        mistral_client=client,
        api_key=mistral_api_key,
        temperature=0.0,
    )
    return model 

Agent code:

agent = Agent(
        name="assistant agent",
        user_id=user_id,
        model=get_mistral_model(),
        ...
)

Actual Behavior

TypeError occured.

Screenshots or Logs (if applicable)

    | Traceback (most recent call last):
    |   File "/home/galaxygliese/.local/lib/python3.10/site-packages/starlette/responses.py", line 260, in wrap
    |     await func()
    |   File "/home/galaxygliese/.local/lib/python3.10/site-packages/starlette/responses.py", line 249, in stream_response
    |     async for chunk in self.body_iterator:
    |   File "/home/galaxygliese/.local/lib/python3.10/site-packages/agno/playground/async_router.py", line 110, in chat_response_streamer
    |     async for run_response_chunk in run_response:
    |   File "/home/galaxygliese/.local/lib/python3.10/site-packages/agno/agent/agent.py", line 968, in _arun
    |     async for model_response_chunk in model_response_stream:  # type: ignore
    | TypeError: 'async for' requires an object with __aiter__ method, got coroutine

Environment

• OS: Ubuntu 22
• Browser (if relevant): Chrome
• Agno Version: agno==1.0.1
• External Dependency Versions: mistralai==1.5.0
• Additional Environment Details: Python 3.10

Possible Solutions (optional)

I think the error may be in async client. The openai version client is shown below:

from openai import AsyncOpenAI as AsyncOpenAIClient
from openai import OpenAI
from agno.models.openai import OpenAIChat 

def get_openai_model()-> OpenAIChat :
    API_KEY = os.environ.get("OPENAI_API_KEY")
    return OpenAIChat(
        id="gpt-4o-mini",
        client=OpenAI(
            api_key=API_KEY,
        ),
        async_client=AsyncOpenAIClient(
            api_key=API_KEY,
        ),
        temperature=0.0,
    )

Additional Context

Add any other context or details about the problem here.

[ashpreetbedi]

Thanks @galaxygliese, looking into this!

[manthanguptaa]

Hey @galaxygliese! Mistral right now doesn't support async capabilities which is causing this issue. Are you by any chance trying to use the playground with Mistral model?

[galaxygliese]

Hi @manthanguptaa! Thank you for your message. Now, I'm trying another model on the playground. No worries! openai-based models are working.

[manthanguptaa]

@galaxygliese we now support async with Mistral in our latest release. Please update agno using pip install -U agno and your code should now start working