BerriAI/litellm · error · ValueError

STABILITY_API_KEY is not set. Please set it via environment

Error message

STABILITY_API_KEY is not set. Please set it via environment variable or pass api_key parameter.

What it means

LiteLLM's Stability image-generation validate_environment resolves the key from the api_key argument or the STABILITY_API_KEY environment variable (get_secret_str). If neither yields a value it raises this ValueError and the generation request never leaves the process.

Source

Thrown at litellm/llms/stability/image_generation/transformation.py:151

        return f"{base_url}{endpoint}"

    def validate_environment(
        self,
        headers: dict,
        model: str,
        messages: list[AllMessageValues],
        optional_params: dict,
        litellm_params: dict,
        api_key: str | None = None,
        api_base: str | None = None,
    ) -> dict:
        """
        Validate environment and set up headers for Stability AI.
        """
        final_api_key: Final[str | None] = api_key or get_secret_str("STABILITY_API_KEY")

        if not final_api_key:
            raise ValueError(
                "STABILITY_API_KEY is not set. Please set it via environment variable or pass api_key parameter."
            )

        headers["Authorization"] = f"Bearer {final_api_key}"
        headers["Accept"] = "application/json"
        return headers

    def transform_image_generation_request(
        self,
        model: str,
        prompt: str,
        optional_params: dict,
        litellm_params: dict,
        headers: dict,
    ) -> dict:
        """
        Transform OpenAI-style request to Stability AI request format.

View on GitHub (pinned to 77b7c6c40c)

Solutions

Set STABILITY_API_KEY in the environment the process actually runs in (export, .env + load_dotenv(), CI/CD secret, container env).
Pass api_key explicitly per call or in the Router/proxy model deployment config.
Confirm the name is exactly STABILITY_API_KEY and is non-empty in that process (os.environ.get).
If using a secrets manager, configure litellm's secret manager rather than relying on plain env vars.

Example fix

# before
resp = litellm.image_generation(model="stability/stable-image-ultra", prompt="a sunset")
# -> ValueError: STABILITY_API_KEY is not set...

# after
import litellm, os
litellm.api_key = None
resp = litellm.image_generation(
    model="stability/stable-image-ultra",
    prompt="a sunset",
    api_key=os.environ["STABILITY_API_KEY"],
)

Defensive patterns

Strategy: validation

Validate before calling

import os


def require_stability_key(api_key: str | None = None) -> str:
    key = api_key or os.environ.get("STABILITY_API_KEY")
    if not key:
        raise RuntimeError(
            "STABILITY_API_KEY missing — set it before generating images"
        )
    return key


key = require_stability_key()
resp = litellm.image_generation(model="stability/stable-image-core", prompt=p, api_key=key)

Try / catch

try:
    resp = litellm.image_generation(model="stability/...", prompt=p)
except ValueError as e:
    if "STABILITY_API_KEY is not set" in str(e):
        raise RuntimeError("deploy config incomplete: add STABILITY_API_KEY") from e
    raise

Prevention

Add env-var presence checks to a startup/healthcheck hook for every provider you enable.
Load .env explicitly (python-dotenv) in dev so behavior matches production secret injection.
Store keys in the platform secret store (K8s secret, Lambda env) instead of shell rc files.

When it happens

Trigger: litellm.image_generation(model="stability/...", ...) with STABILITY_API_KEY unset/empty in the process environment and no api_key kwarg; serverless deployments where the env var was set in a different stage; notebooks after a kernel restart that lost os.environ changes.

Common situations: Local works / production fails because the key was only set in the developer shell; secrets managers (Vault, AWS Secrets Manager) not wired into the deployment; a .env file present but python-dotenv load_dotenv() never called; proxy restarted without the env var.

Understand the failure class

Background: "API key is required" / "API key not found" / "No API key was set": the missing-api-key error family across 16 libraries — this error's family across 16 libraries.

Related errors

AI-assisted analysis of BerriAI/litellm@77b7c6c40c (2026-08-18). Data as JSON: /api/errors/ca0455b73c23dc06. Report an issue: GitHub.