chroma-core/chroma · error · ValueError

Vertex AI and API key are mutually exclusive in the client i

Error message

Vertex AI and API key are mutually exclusive in the client initializer.

What it means

Raised in GoogleGeminiEmbeddingFunction.__init__ when the environment variable named by api_key_env_var (default GEMINI_API_KEY) is set AND vertexai is truthy. The underlying genai.Client authenticates either with an API key or with Vertex AI Application Default Credentials, never both; Chroma fails fast instead of letting the SDK silently pick one auth path.

Source

Thrown at chromadb/utils/embedding_functions/google_embedding_function.py:59

            location (str, optional): The Google Cloud location/region (required for Vertex AI).
        """
        try:
            import google.genai as genai
        except ImportError:
            raise ValueError(
                "The google-genai python package is not installed. Please install it with `pip install google-genai`"
            )

        self.model_name = model_name
        self.task_type = task_type
        self.dimension = dimension
        self.api_key_env_var = api_key_env_var
        self.vertexai = vertexai
        self.project = project
        self.location = location
        self.api_key = os.getenv(self.api_key_env_var) if self.api_key_env_var else None
        if self.api_key and self.vertexai:
            raise ValueError(
                "Vertex AI and API key are mutually exclusive in the client initializer."
            )
        if not self.api_key and not self.vertexai:
            raise ValueError(
                f"The {self.api_key_env_var} environment variable must be set if vertexai is not enabled."
            )

        from google.genai import types

        self.client = genai.Client(
            api_key=self.api_key,
            vertexai=vertexai,
            project=project,
            location=location,
            http_options=types.HttpOptions(
                headers={"x-goog-api-client": f"chroma/{__version__}"}
            ),
        )

View on GitHub (pinned to aecdd12c8a)

Solutions

1. Unset the key in the environment that uses Vertex: unset GEMINI_API_KEY (or remove it from .env / compose env)
2. Or point api_key_env_var at a variable that is deliberately unset in that deployment
3. Or drop vertexai (leave it None/False) to keep plain API-key auth
4. If you genuinely need both auth modes, run them in separate processes/environments with distinct env var sets

Example fix

# before
# export GEMINI_API_KEY=...  (still set in the shell)
ef = GoogleGeminiEmbeddingFunction(
    vertexai=True, project="my-gcp-project", location="us-central1"
)  # ValueError: mutually exclusive

# after
# unset GEMINI_API_KEY
ef = GoogleGeminiEmbeddingFunction(
    vertexai=True, project="my-gcp-project", location="us-central1"
)

import os

API_KEY_ENV = "GEMINI_API_KEY"
use_vertex = True

if use_vertex and os.getenv(API_KEY_ENV):
    raise SystemExit(
        f"{API_KEY_ENV} must not be set when vertexai=True; unset it or use API-key auth"
    )

When it happens

Trigger: Constructing GoogleGeminiEmbeddingFunction(vertexai=True, project=..., location=...) while GEMINI_API_KEY (or whatever var api_key_env_var points to) is present in the environment; or passing an api_key_env_var that names a variable that happens to be set in the Vertex deployment.

Common situations: Migrating an app from the Gemini API to Vertex AI without unsetting the old key; a globally loaded .env or shell profile exporting GEMINI_API_KEY into a production Vertex container; CI secrets injecting the key into a job that also sets vertexai=True.

Related errors

• The {self.api_key_env_var} environment variable must be set
• The vertexai cannot be changed after the embedding function
• Google API key is required. Please provide it in the constru
• Your API key does not have access to any DBs for tenant ${th
• Unauthorized

AI-assisted analysis of chroma-core/chroma@aecdd12c8a (2026-08-16). Data as JSON: /api/errors/58cc830589653178. Report an issue: GitHub.