BerriAI/litellm · error · ValueError

api_base is required in generic_prompt_config

Error message

api_base is required in generic_prompt_config

What it means

prompt_initializer raises ValueError when litellm_params.api_base is falsy while initializing a prompt from a generic prompt-management API. Generic prompt management fetches prompts over HTTP from a user-supplied base URL, so api_base is mandatory whenever a prompt_id/prompt_spec is used with a generic provider.

Solutions

  1. Add api_base to your generic_prompt_config, e.g. api_base: https://prompts.internal.example.com.
  2. Confirm it sits at the level the code reads (litellm_params of the prompt config), not under a nested provider block.
  3. Also set api_key if the prompt API requires auth.
  4. Check for empty-string api_base from templated configs — falsy values fail.

Example fix

# before
litellm.global_generic_prompt_config = {'api_key': 'sk-...'}

# after
litellm.global_generic_prompt_config = {'api_base': 'https://prompts.internal.example.com', 'api_key': 'sk-...'}
Defensive patterns

Strategy: validation

Validate before calling

if not litellm_params.api_base:
    raise ValueError('generic_prompt_config requires api_base (the prompt API URL)')

Type guard

def prompt_config_has_api_base(prompt_config: dict) -> bool:
    return bool(prompt_config.get('api_base'))

Try / catch

try:
    prompt_initializer(litellm_params, prompt_spec)
except ValueError as e:
    if 'api_base' in str(e):
        raise ConfigError('Add api_base to generic_prompt_config') from e
    raise

Prevention

When it happens

Trigger: Routing a completion with a prompt_id and a provider marked as generic prompt management while prompt_config/generic_prompt_config lacks api_base; setting api_key but not api_base in the prompt config; the api_base key nested in the wrong section of config.yaml.

Common situations: Wiring LiteLLM to an internal prompt-management service and forgetting the URL; copy-pasting a generic_prompt_config example that only shows api_key; config structured for a built-in prompt provider (which needs no api_base) then switching to generic without adding it.

Understand the failure class

Background: "X is required", "must be set", "cannot be empty": the missing-required-config error family, from Vertex AI project/location to WeChat keys — this error's family across 18 libraries.

Related errors

AI-assisted analysis of BerriAI/litellm@77b7c6c40c (2026-08-18). Data as JSON: /api/errors/80126977ac105c3c. Report an issue: GitHub.

Appendix: source

Thrown at litellm/integrations/generic_prompt_management/__init__.py:43

                - api_base: Base URL for the API
                - api_key: Optional API key for authentication
                - timeout: Request timeout in seconds (default: 30)
    """
    import litellm

    litellm.global_generic_prompt_config = config


def prompt_initializer(litellm_params: "PromptLiteLLMParams", prompt_spec: "PromptSpec") -> "CustomPromptManagement":
    """
    Initialize a prompt from a generic prompt management API.
    """
    prompt_id: Final = getattr(litellm_params, "prompt_id", None)

    api_base: Final = litellm_params.api_base
    api_key: Final = litellm_params.api_key
    if not api_base:
        raise ValueError("api_base is required in generic_prompt_config")

    provider_specific_query_params: Final = litellm_params.provider_specific_query_params

    try:
        generic_prompt_manager: Final = GenericPromptManager(
            api_base=api_base,
            api_key=api_key,
            prompt_id=prompt_id,
            additional_provider_specific_query_params=provider_specific_query_params,
            **litellm_params.model_dump(
                exclude_none=True,
                exclude={
                    "prompt_id",
                    "api_key",
                    "provider_specific_query_params",
                    "api_base",
                },
            ),

View on GitHub (pinned to 77b7c6c40c)