Configuration

This guide covers all configuration options for the EndeeVectorStore class.

Constructor Parameters

The EndeeVectorStore constructor accepts the following parameters:

Parameter	Type	Required	Description
`type`	`str`	Yes	Name of the Endee index
`api_token`	`str`	Yes*	Your Endee API token (required if `endee_index` not provided)
`embedder_config`	`dict`	Yes	Configuration for the embedding provider
`space_type`	`str`	No	Distance metric: `"cosine"`, `"l2"`, or `"ip"` (default: `"cosine"`)
`allow_reset`	`bool`	No	Whether to allow resetting the index (default: `True`)
`crew`	`Crew`	No	Optional CrewAI object for integration
`text_key`	`str`	No	Key to store text in metadata (default: `"value"`)
`endee_index`	`Index`	No	Optional existing Endee index object

Distance Metrics

Metric	Best For
`cosine`	Text embeddings, normalized vectors (default)
`l2`	Image features, spatial data
`ip`	Recommendation systems, dot product similarity

Embedder Configuration

The embedder_config parameter configures your embedding provider:


embedder_config = {
    "provider": "<provider_name>",
    "config": {
        "model_name": "<model_name>",
        "api_key": "<api_key>"
    }
}

Supported Providers

OpenAI:


embedder_config = {
    "provider": "openai",
    "config": {
        "model_name": "text-embedding-3-small",
        "api_key": "<OPENAI_API_KEY>"
    }
}

Google:


embedder_config = {
    "provider": "google",
    "config": {
        "model_name": "models/embedding-001",
        "api_key": "<GOOGLE_API_KEY>"
    }
}

Cohere:


embedder_config = {
    "provider": "cohere",
    "config": {
        "model_name": "small",
        "api_key": "<COHERE_API_KEY>"
    }
}

HuggingFace:


embedder_config = {
    "provider": "huggingface",
    "config": {
        "model_name": "sentence-transformers/all-MiniLM-L6-v2"
    }
}

Full Configuration Example


from endee_crewai import EndeeVectorStore
 
memory_store = EndeeVectorStore(
    type="my_knowledge_base",
    api_token="<ENDEE_API_TOKEN>",
    embedder_config={
        "provider": "openai",
        "config": {
            "model_name": "text-embedding-3-small",
            "api_key": "<OPENAI_API_KEY>"
        }
    },
    space_type="cosine",
    allow_reset=True,
    crew=None,
    text_key="value"
)

Methods Reference

EndeeVectorStore Methods

Method	Description
`save(text, metadata)`	Save a document with metadata
`search(query, limit, filter)`	Search for similar documents
`reset()`	Delete and recreate the index

Save Method


memory_store.save(
    text="Document content here",
    metadata={"key": "value", "category": "example"}
)

Search Method


results = memory_store.search(
    query="search query",
    limit=10,
    filter={"category": {"$eq": "example"}}
)

Reset Method


# Reset the index (deletes all data)
memory_store.reset()
 
# Wait for reset to complete
import time
time.sleep(2)

Warning: The reset() method permanently deletes all data in the index. Use with caution.

Environment Variables

For production deployments, use environment variables instead of hardcoding API keys:


import os
 
memory_store = EndeeVectorStore(
    type="production_index",
    api_token=os.environ["ENDEE_API_TOKEN"],
    embedder_config={
        "provider": "openai",
        "config": {
            "model_name": "text-embedding-3-small",
            "api_key": os.environ["OPENAI_API_KEY"]
        }
    },
    space_type="cosine"
)