BerriAI/litellm · error · Exception

PromptInjectionDetection: Invalid LLM API Name. LLM API Name

Error message

PromptInjectionDetection: Invalid LLM API Name. LLM API Name must be a 'model_name' in 'model_list'.

What it means

Error "PromptInjectionDetection: Invalid LLM API Name. LLM API Name must be a 'model_name' in 'model_list'." thrown in BerriAI/litellm.

Source

Thrown at litellm/proxy/hooks/prompt_injection_detection.py:93

            print(print_statement)  # noqa: T201

    def update_environment(self, router: Router | None = None):
        self.llm_router = router

        if self.prompt_injection_params is not None and self.prompt_injection_params.llm_api_check is True:
            if self.llm_router is None:
                raise Exception(
                    "PromptInjectionDetection: Model List not set. Required for Prompt Injection detection."
                )

            self.print_verbose(
                f"model_names: {self.llm_router.model_names}; self.prompt_injection_params.llm_api_name: {self.prompt_injection_params.llm_api_name}"
            )
            if (
                self.prompt_injection_params.llm_api_name is None
                or self.prompt_injection_params.llm_api_name not in self.llm_router.model_names
            ):
                raise Exception(
                    "PromptInjectionDetection: Invalid LLM API Name. LLM API Name must be a 'model_name' in 'model_list'."
                )

    def generate_injection_keywords(self) -> list[str]:
        combinations: Final = []
        for verb in self.verbs:
            for adj in self.adjectives:
                for prep in self.prepositions:
                    phrase = " ".join(filter(None, [verb, adj, prep])).strip()
                    if len(phrase.split()) > 2:  # additional check to ensure more than 2 words
                        combinations.append(phrase.lower())
        return combinations

    def check_user_input_similarity(
        self,
        user_input: str,
        similarity_threshold: float = DEFAULT_PROMPT_INJECTION_SIMILARITY_THRESHOLD,
    ) -> bool:

View on GitHub (pinned to 77b7c6c40c)

Solutions

1. Set llm_api_name to a model_name that exists in the proxy model_list.
2. Check for typos in the configured LLM API name.

When it happens

Trigger: Thrown at litellm/proxy/hooks/prompt_injection_detection.py:93 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/28738b5c3bd60539. Report an issue: GitHub.