Skip to main content
Version: V1.1.0

Cohere

Cohere provides embedding models for semantic search, clustering, and recommendation. seekdb provides a CohereEmbeddingFunction wrapper (powered by LiteLLM) to generate Cohere embeddings and use them with seekdb collections.

tip

Using Cohere service requires you to follow Cohere's pricing rules and may incur corresponding fees. Before proceeding, please visit their official website or refer to relevant documentation to confirm and accept their pricing standards. If you do not agree, please do not proceed.

Dependencies and authentication​

CohereEmbeddingFunction calls the Cohere embedding API via LiteLLM. In practice, you typically need:

  • A Cohere API key with access to the embedding models you plan to use
  • Python packages: pyseekdb and litellm

Authentication is usually provided via environment variables (by default, COHERE_API_KEY). If you use a different environment variable name, set api_key_env.

Example: create a Cohere embedding function​

  • Basic usage

    from pyseekdb.utils.embedding_functions import CohereEmbeddingFunction

    ef = CohereEmbeddingFunction(model_name="embed-english-v3.0")

  • Multilingual model

    from pyseekdb.utils.embedding_functions import CohereEmbeddingFunction

    ef = CohereEmbeddingFunction(model_name="embed-multilingual-v3.0")

  • Use a custom API key environment variable (api_key_env)

    from pyseekdb.utils.embedding_functions import CohereEmbeddingFunction

    ef = CohereEmbeddingFunction(
        model_name="embed-english-v3.0",
        api_key_env="YOUR_API_KEY_ENV",
    )

  • Set input_type (recommended for retrieval)

    from pyseekdb.utils.embedding_functions import CohereEmbeddingFunction

    ef = CohereEmbeddingFunction(
        model_name="embed-english-v3.0",
        input_type="search_document",
    )

    # For query-side embeddings, use input_type="search_query"

Parameters

  • model_name: Cohere embedding model name (for example, embed-english-v3.0).
  • api_key_env: the environment variable name holding your Cohere API key (default: COHERE_API_KEY).
  • input_type: an optional hint to optimize retrieval embeddings (commonly search_document or search_query).

References​

Contents

Edit this page