BerriAI/litellm · error · ValueError

embedding_config is required in litellm_params for Azure AI

Error message

embedding_config is required in litellm_params for Azure AI Search. Example: litellm_params['embedding_config'] = {'api_base': 'https://krris-mh44uf7y-eastus2.cognitiveservices.azure.com/', 'api_key': 'os.environ/AZURE_API_KEY', 'api_version': '2025-09-01'}

What it means

Alongside the embedding model, the search transform requires the connection settings for it. litellm_params.get('litellm_embedding_config', {}) returns an empty dict by default, and an empty dict is falsy, so both a missing key and an explicitly empty dict raise this error.

Source

Thrown at litellm/llms/azure_ai/vector_stores/transformation.py:138

        Transform search request for Azure AI Search API

        Generates embeddings using litellm.embeddings and constructs Azure AI Search request
        """
        # Convert query to string if it's a list
        if isinstance(query, list):
            query = " ".join(query)

        # Get embedding model from litellm_params (required)
        embedding_model: Final = litellm_params.get("litellm_embedding_model")
        if not embedding_model:
            raise ValueError(
                "embedding_model is required in litellm_params for Azure AI Search. "
                "Example: litellm_params['embedding_model'] = 'azure/text-embedding-3-large'"
            )

        embedding_config: Final = litellm_params.get("litellm_embedding_config", {})
        if not embedding_config:
            raise ValueError(
                "embedding_config is required in litellm_params for Azure AI Search. "
                "Example: litellm_params['embedding_config'] = {'api_base': 'https://krris-mh44uf7y-eastus2.cognitiveservices.azure.com/', 'api_key': 'os.environ/AZURE_API_KEY', 'api_version': '2025-09-01'}"
            )

        # Get vector field name (defaults to contentVector)
        vector_field: Final = litellm_params.get("azure_search_vector_field", "contentVector")

        # Get top_k (number of results to return)
        top_k: Final = vector_store_search_optional_params.get("top_k", 10)

        # Generate embedding for the query using litellm.embeddings
        try:
            embedding_response: Final = litellm.embedding(
                model=embedding_model,
                input=[query],
                **embedding_config,
            )
            query_vector: Final = embedding_response.data[0]["embedding"]

View on GitHub (pinned to 6c2dcb801b)

Solutions

1. Set litellm_params['litellm_embedding_config'] = {'api_base': 'https://<resource>.cognitiveservices.azure.com/', 'api_key': os.environ['AZURE_API_KEY'], 'api_version': '2025-09-01'}
2. Note the key prefix: it is litellm_embedding_config, not embedding_config
3. If the embedding deployment lives in the same resource as OCR/chat, reuse those credentials explicitly — they are not inherited

Example fix

# before
litellm_params={..., 'litellm_embedding_model': 'azure/text-embedding-3-large'}  # no config

# after
litellm_params={..., 'litellm_embedding_model': 'azure/text-embedding-3-large', 'litellm_embedding_config': {'api_base': 'https://my-resource.cognitiveservices.azure.com/', 'api_key': os.environ['AZURE_API_KEY'], 'api_version': '2025-09-01'}}

emb_cfg = litellm_params.get('litellm_embedding_config')
if not emb_cfg or not isinstance(emb_cfg, dict):
    raise ValueError('litellm_embedding_config must be a non-empty dict with api_base/api_key')

def is_valid_embedding_config(cfg) -> bool:
    return isinstance(cfg, dict) and bool(cfg.get('api_base')) and bool(cfg.get('api_key'))

When it happens

Trigger: Configuring litellm_embedding_model but omitting litellm_embedding_config; or setting it to {} — both fail. The config must contain at minimum the api_base and api_key (and typically api_version) for the embedding deployment.

Common situations: Assuming the embedding call reuses the vector store's api_key; supplying the config under embedding_config without the litellm_ prefix; leaving a placeholder {} while migrating.

Related errors

• embedding_model is required in litellm_params for Azure AI S
• Azure AI Search service name is required. Provide it via lit
• api_base is None. Please set AZURE_AI_API_BASE or dynamicall
• api_key is None. Please set AZURE_AI_API_KEY or dynamically
• api_key is required

AI-assisted analysis of BerriAI/litellm@6c2dcb801b (2026-08-15). Data as JSON: /api/errors/935f49b91fa75822. Report an issue: GitHub.