langchain-cyborgdb

LangChain integration for CyborgDB, the Confidential Vector Database. Provides CyborgVectorStore, a LangChain VectorStore backed by CyborgDB's encrypted vector indexes.

This package is a thin wrapper around the integration that ships in cyborgdb-core — it re-exports CyborgVectorStore, CyborgEmbeddings, and StorageConfig under the standalone package name that LangChain's integration directory expects.

Installation

pip install langchain-cyborgdb

Quick Start

The example below uses OpenAI embeddings (pip install langchain-openai), but any LangChain Embeddings implementation works. You'll need a CyborgDB API key from cyborg.co.

from langchain_cyborgdb import CyborgVectorStore, StorageConfig
from langchain_openai import OpenAIEmbeddings

store = CyborgVectorStore.from_texts(
    texts=["hello world", "goodbye world"],
    embedding=OpenAIEmbeddings(),
    index_key=CyborgVectorStore.generate_key(),
    api_key="your-api-key",
    storage_config=StorageConfig.disk("/tmp/cyborgdb-langchain"),
    metric="cosine"
)

docs = store.similarity_search("hello")

Key Features

• Full VectorStore API compatibility (add_texts, similarity_search, etc.)
• Async operations support
• Configurable distance metrics (cosine, euclidean, squared_euclidean)
• Document ID tracking and management
• Max marginal relevance search
• Factory methods (from_texts, from_documents) for quick setup

Key Classes

CyborgVectorStore

Implements the LangChain-compatible vector store interface. Supports:

• .add_texts() / .add_documents()
• .similarity_search() / .similarity_search_with_score()
• .max_marginal_relevance_search()
• .get_document() / .get_documents() / .list_ids() / .delete()
• .as_retriever() for use in LangChain chains
• Async variants (.aadd_texts(), .asimilarity_search(), etc.)

StorageConfig

Configuration object for defining where index data is stored, built with StorageConfig.disk(path) or StorageConfig.s3(bucket, ...).

Documentation

Full documentation, including metadata filtering and hybrid (BM25) search, lives at docs.cyborg.co.

Using the CyborgDB client/server deployment instead of the embedded engine? The same integration is available via pip install cyborgdb[langchain] — see the Deployment Models Guide.

License

This wrapper is MIT licensed. It depends on cyborgdb-core, which is distributed under its own proprietary license.