Welcome to LangChain MatrixOne’s documentation!

This guide explains how to connect LangChain applications to MatrixOne using the MatrixOneVectorStore integration included in this package.

Prerequisites

  • A running MatrixOne deployment accessible via the MySQL protocol.

  • Credentials with privileges to create tables and insert/query data.

  • A LangChain-compatible embedding model (OpenAI, Hugging Face, Azure, etc.).

Installation

Install the package from PyPI with your preferred tool:

pip install langchain-matrixone
# or
uv pip install langchain-matrixone

Quickstart Example

from langchain_matrixone import MatrixOneVectorStore
from langchain_community.embeddings import HuggingFaceEmbeddings
from langchain_text_splitters import RecursiveCharacterTextSplitter
from langchain_core.documents import Document

 # make sure you have a running MatrixOne deployment accessible via the MySQL protocol.
connection_args = {
    "host": "127.0.0.1",
    "port": 6001,
    "user": "root",
    "password": "111",
    "database": "langchain_demo",
}

embeddings = HuggingFaceEmbeddings(model_name="sentence-transformers/all-MiniLM-L6-v2")
vector_store = MatrixOneVectorStore(
    embedding=embeddings,
    connection_args=connection_args,
    table_name="langchain_vectors",
)

docs = RecursiveCharacterTextSplitter(chunk_size=400, chunk_overlap=40).split_documents(
    [
        Document(page_content="MatrixOne offers fast vector search with SQL semantics."),
        Document(page_content="LangChain unifies LLM tooling in Python and JavaScript."),
    ]
)

vector_store.add_texts(
    [doc.page_content for doc in docs],
    metadatas=[doc.metadata for doc in docs],
)

results = vector_store.similarity_search("What does LangChain provide?", k=2)
for doc in results:
    print(doc.page_content, doc.metadata)

Reusing an existing client

You can pass a pre-configured matrixone.Client if your application already maintains pooled connections:

from matrixone import Client

client = Client()
client.connect(**connection_args)

vector_store = MatrixOneVectorStore(
    embedding=embeddings,
    client=client,
    table_name="langchain_vectors",
)

Notes

  • Tables are created automatically on first use; pass drop_old=True to rebuild.

  • Metadata is stored as JSON in MatrixOne—ensure values are serializable.

  • Keep the embedding dimension consistent per table to avoid schema mismatches.

  • See the MatrixOne Python SDK documentation for advanced vector index management and connection best practices.

Contents:

Indices and tables