langchain-astradb

What it is and what it does

langchain-astradb is a LangChain integration package that connects to DataStax Astra DB, a serverless vector-capable database built on Apache Cassandra. It replaces deprecated community-maintained Astra DB classes and provides six main components: AstraDBVectorStore for vector similarity search (including hybrid search with BM25 and reranking), AstraDBChatMessageHistory for storing conversation state, AstraDBCache and AstraDBSemanticCache for LLM response caching, AstraDBLoader for retrieving documents, and AstraDBStore/AstraDBByteStore for key-value storage.

The package depends on langchain-core, astrapy, and numpy. It handles collection creation idempotently by default, meaning it will create the underlying Astra DB collection on first use if it doesn't exist, or verify configuration matches if it does. The documentation emphasizes configuration matching between your code and the database to avoid mismatch errors, particularly around hybrid search settings and indexing options. It supports autodetection of collection configuration and offers explicit setup modes for predictable behavior.

Use it for:

• Build RAG (retrieval-augmented generation) applications using AstraDBVectorStore for semantic search over document embeddings.
• Store and retrieve multi-turn conversation history in stateless LLM applications via AstraDBChatMessageHistory.
• Cache LLM responses to reduce API costs and latency using AstraDBCache or AstraDBSemanticCache for semantic deduplication.
• Load and index documents from Astra DB collections into LangChain workflows using AstraDBLoader.
• Implement hybrid search combining vector similarity, keyword matching (BM25), and reranking in a single query.
• Migrate from deprecated langchain_community Astra DB classes to get latest features and modern astrapy compatibility.

Worth the install?

AI-flagged interpretation of the facts on this page — verify before relying

Integrates Astra DB (a serverless vector database built on Apache Cassandra) with LangChain, providing vector storage, chat history, caching, and document loading capabilities.

Yes. This is an actively maintained, low-friction integration with no known vulnerabilities and permissive licensing. Install it if you are building LangChain applications on Astra DB or migrating away from deprecated community plugins. Be aware that collection configuration must match your code's expectations to avoid mismatch errors, particularly around hybrid search and indexing settings.

Install

langchain-astradb on PyPI

pip install langchain-astradb

uv add langchain-astradb

poetry add langchain-astradb

Installing langchain-astradb

Before you install

Low install friction with a pure-Python wheel distribution. Actively maintained with recent commits and no known vulnerabilities. Requires Python 3.10 or later.

License in practice

MIT license permits free use, modification, and distribution with minimal restrictions—suitable for both open-source and commercial projects.

Quickstart

pip install langchain-astradb

from langchain_astradb import AstraDBVectorStore

vector_store = AstraDBVectorStore(
    embedding=my_embedding,
    collection_name="my_collection",
    api_endpoint="https://...",
    token="AstraCS:...",
)

Requires an active Astra DB instance with API endpoint and application token; Python 3.10 or later.

Verify before relying

• Whether server-side embedding computation ('vectorize') requires additional setup or API keys beyond the Astra DB token.
• Performance characteristics and scaling limits for the vector store with large document collections.
• Compatibility guarantees with specific versions of langchain-core and astrapy beyond what the dependency pins enforce.

Package facts

Evidence: langchain_astradb-1.0.0-py3-none-any.whl

Tags

More Artificial Intelligence packages