BerriAI/litellm · error · HTTPException

LLM Router not initialized. Ensure models are added to proxy

Error message

LLM Router not initialized. Ensure models are added to proxy.

What it means

llm_router is None while routing a managed vector-store request: the proxy has no models configured, so there is no router to resolve the target model names. Add models via config or the /model/new endpoint before using multi-model vector store support.

Source

Thrown at litellm/proxy/vector_store_endpoints/endpoints.py:232

            target_model_names_list = [m.strip() for m in target_model_names.split(",")]
        elif isinstance(target_model_names, list):
            target_model_names_list = target_model_names
        else:
            raise HTTPException(
                status_code=400,
                detail="target_model_names must be a comma-separated string or list of model names",
            )

        # Get managed vector stores hook
        managed_vector_stores: Final[Any] = proxy_logging_obj.get_proxy_hook("managed_vector_stores")
        if managed_vector_stores is None:
            raise HTTPException(
                status_code=500,
                detail="Managed vector stores not configured. Please ensure the proxy is initialized with database support.",
            )

        if llm_router is None:
            raise HTTPException(
                status_code=500,
                detail="LLM Router not initialized. Ensure models are added to proxy.",
            )

        # Create vector store across multiple models
        response: Final = await managed_vector_stores.acreate_vector_store(
            create_request=data,
            llm_router=llm_router,
            target_model_names_list=target_model_names_list,
            litellm_parent_otel_span=user_api_key_dict.parent_otel_span,
            user_api_key_dict=user_api_key_dict,
        )

        return response

    processor: Final = ProxyBaseLLMRequestProcessing(data=data)
    try:
        return await processor.base_process_llm_request(

View on GitHub (pinned to 77b7c6c40c)

Solutions

1. Add models to the proxy (model_list in config.yaml) so the LLM router initializes.

When it happens

Trigger: Thrown at litellm/proxy/vector_store_endpoints/endpoints.py:232 when the library encounters an invalid state.

Common situations: See trigger scenarios.

AI-assisted analysis of BerriAI/litellm@77b7c6c40c (2026-08-18). Data as JSON: /api/errors/6f68de2ec935b27b. Report an issue: GitHub.