BerriAI/litellm · error · ValueError

Router not initialized

Error message

Router not initialized

What it means

Runtime guard when executing a guardrail via the router: llm_router on proxy_server is None, meaning no models (and hence no router) were configured — typically the proxy started with an empty config or the router failed to initialize. Guardrail resolution cannot proceed without it.

Source

Thrown at litellm/proxy/utils.py:1065

    ) -> object:
        """
        Execute a guardrail using the router's load balancing.

        Args:
            guardrail_name: Name of the guardrail
            hook_type: One of "pre_call", "during_call", "post_call"
            data: Request data
            user_api_key_dict: User API key auth
            call_type: Type of call
            response: Response object (for post_call hooks)

        Returns:
            Result from the guardrail execution
        """
        from litellm.proxy.proxy_server import llm_router

        if llm_router is None:
            raise ValueError("Router not initialized")

        # Select guardrail using router's load balancing
        selected_guardrail: Final = llm_router.get_available_guardrail(guardrail_name=guardrail_name)

        callback: Final[CustomGuardrail | None] = selected_guardrail.get("callback")
        if callback is None:
            raise ValueError(f"No callback found for guardrail: {guardrail_name}")

        return await self._execute_guardrail_hook(
            callback=callback,
            hook_type=hook_type,
            data=data,
            user_api_key_dict=user_api_key_dict,
            call_type=call_type,
            response=response,
        )

    async def _process_guardrail_callback(

View on GitHub (pinned to 77b7c6c40c)

Solutions

1. Ensure the proxy/router finished initializing before invoking this code path; check startup logs for errors.

When it happens

Trigger: Thrown at litellm/proxy/utils.py:1065 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/6f2e7f28ae98460b. Report an issue: GitHub.