Using ShortTermMemory with Azure

[arashaga]

I am geting an error when I trying to use Azure open ai using shorttermmemory :
ValidationError: 1 validation error for Crew
Value error, Please provide an OpenAI API key. You can get one at https://platform.openai.com/account/api-keys [type=value_error, input_value={'agents': [Agent(role=Re...at 0x0000022E277CE5D0>)}, input_type=dict]
For further information visit https://errors.pydantic.dev/2.10/v/value_error

this is my code:

# === Crew Orchestration ===
from crewai.memory.storage.rag_storage import RAGStorage
from crewai.memory.short_term.short_term_memory import ShortTermMemory
embedder_config={
    "provider": "azure",
    "config": {
                "api_key": AZURE_OPENAI_API_KEY,
                "api_base": AZURE_OPENAI_ENDPOINT,
                "api_version": "2024-02-01",
                "model_name": 'text-embedding-3-small'
    }

}
crew = Crew(
    agents=[search_agent, specialization_agent],
    tasks=[search_task,specialist_task],
    process=Process.sequential,
    verbose=True,
    memory=True,
    short_term_memory = ShortTermMemory(
    storage=RAGStorage(
        embedder_config=embedder_config,
        type="short_term",
        path="/my_crew1/"
    )
)
)

Can someone please let me know what am I missing?

[Vidit-Ostwal]

Can you pass a llm varaible to the crew as well?
Default is set to OPENAI, that's why the validation error?

[arashaga]

I did and got the same error:

from crewai.memory.storage.rag_storage import RAGStorage
from crewai.memory.short_term.short_term_memory import ShortTermMemory
embedder_config={
    "provider": "azure",
    "config": {
                "api_key": AZURE_OPENAI_API_KEY,
                "api_base": AZURE_OPENAI_ENDPOINT,
                "api_version": "2024-02-01",
                "model_name": 'text-embedding-3-small'
    }

}
crew = Crew(
    agents=[search_agent, specialization_agent,report_builder_agent],
    tasks=[search_task,specialist_task,report_builder_task],
    process=Process.sequential,
    verbose=True,
    llm=azure_llm,
   memory=True,
    short_term_memory = ShortTermMemory(
    storage=RAGStorage(
        embedder_config=embedder_config,
        type="short_term",
        path="/my_crew1/"
    )
)
)

FYI my crewai version is 0.108.0

[Vidit-Ostwal]

Can you once try passing the llm at the agent level as well?

[arashaga]

it's already passed to them.

[arashaga]

any updates on this?

[arashaga]

bumping this up again. Any updates on this?

[Vidit-Ostwal]

@arashaga, I think the issue is with the Chromadb side, there is one another issue some other user had raised
#2655

Do let me know if this helps.
If possible can you try a different provider